<%BANNER%>

Effects of Tropical Cyclones on Circulation and Momentum Balance in a Subtropical Estuary and Inlet

Permanent Link: http://ufdc.ufl.edu/UFE0042341/00001

Material Information

Title: Effects of Tropical Cyclones on Circulation and Momentum Balance in a Subtropical Estuary and Inlet
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Tutak, Bilge
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: ch3d, guana, hurricane, hydrodynamic, momentum, northern, residual, salinity, st, tropical
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

Abstract: Estuaries along the coast of U.S. are subjected to extreme events such as tropical cyclones, nor easters, and heavy rainfall. These extreme events can bring very drastic and sudden changes to the estuarine environment and habitat. This study focuses on the hydrodynamics of the subtropical Northern Coastal Basin (NCB) estuary during regular (non-storm) time periods as well as tropical cyclones. Hydrodynamics (water level, currents and salinity) of the system is simulated with a three-dimensional model, CH3D (Curvilinear-grid Hydrodynamics in 3D), during four major hurricanes (Charley, Frances, Ivan, Jeanne) in 2004 and two tropical storms (Fay and Hanna) in 2008. After successful calibration and verification of using the observed water level, salinity and current (2008) data throughout the estuary, the model is used to examine the residual circulation, momentum balance, and stratification in order to understand the effects of tropical cyclones on the estuarine hydrodynamics. Model results during a typical tidal cycle show formation of four residual eddies around the headlands of the St. Augustine Inlet due to the nonlinear advection. The dynamic structure of the residual eddies strongly influences the cross-sectional residual flow inside the inlet. Analysis of the residual momentum balance simulated by the model reveals that during a typical tidal cycle with weak external forcing, the balance is between the barotropic pressure gradient and the nonlinear advection. Prior to the landfall of storms, strong northeasterly winds disrupted the ocean side eddies and created a uniform along-shore southerly current which forced complete inflow through the inlet due to Ekman transport, but eddies inside the inlet persisted. After the landfall, the wind direction changed to easterly followed by southeasterly winds. The change in wind direction caused the along-shore current to completely reverse and become northerly on the ocean side, while the eddies inside the inlet disrupted and the flow through inlet became complete outflow. The duration of the changes were dependent on the duration of the storm that affected the region, ranging from 3-4 days for Charley and Fay, and up to 12 days for Frances. The momentum balance within the system also changed significantly during storms. Wind stress generated by strong winds was mostly balanced by a combination of barotropic pressure gradient and bottom friction. Baroclinic pressure gradient became important coinciding with increased precipitation or river discharge events in the system and it was more significant in 2004 compared to the 2008 simulations due to the higher precipitation and river discharge (3 times higher). Salinity inside the estuary was significantly affected by the increased precipitation and river discharge especially during the storms. When the residual flow rate at the inlets was directed towards the ocean, the size of low salinity area was at its maximum. Although the salinity within the system was highly affected by precipitation and river discharge, the stratification within the system was generally very weak due to the shallow and dynamic nature of the system except the regions closer to the rivers and creeks. The model results show two potential regions with a very slow flushing rate.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Bilge Tutak.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Sheng, Y. P.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042341:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042341/00001

Material Information

Title: Effects of Tropical Cyclones on Circulation and Momentum Balance in a Subtropical Estuary and Inlet
Physical Description: 1 online resource (118 p.)
Language: english
Creator: Tutak, Bilge
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2010

Subjects

Subjects / Keywords: ch3d, guana, hurricane, hydrodynamic, momentum, northern, residual, salinity, st, tropical
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

Abstract: Estuaries along the coast of U.S. are subjected to extreme events such as tropical cyclones, nor easters, and heavy rainfall. These extreme events can bring very drastic and sudden changes to the estuarine environment and habitat. This study focuses on the hydrodynamics of the subtropical Northern Coastal Basin (NCB) estuary during regular (non-storm) time periods as well as tropical cyclones. Hydrodynamics (water level, currents and salinity) of the system is simulated with a three-dimensional model, CH3D (Curvilinear-grid Hydrodynamics in 3D), during four major hurricanes (Charley, Frances, Ivan, Jeanne) in 2004 and two tropical storms (Fay and Hanna) in 2008. After successful calibration and verification of using the observed water level, salinity and current (2008) data throughout the estuary, the model is used to examine the residual circulation, momentum balance, and stratification in order to understand the effects of tropical cyclones on the estuarine hydrodynamics. Model results during a typical tidal cycle show formation of four residual eddies around the headlands of the St. Augustine Inlet due to the nonlinear advection. The dynamic structure of the residual eddies strongly influences the cross-sectional residual flow inside the inlet. Analysis of the residual momentum balance simulated by the model reveals that during a typical tidal cycle with weak external forcing, the balance is between the barotropic pressure gradient and the nonlinear advection. Prior to the landfall of storms, strong northeasterly winds disrupted the ocean side eddies and created a uniform along-shore southerly current which forced complete inflow through the inlet due to Ekman transport, but eddies inside the inlet persisted. After the landfall, the wind direction changed to easterly followed by southeasterly winds. The change in wind direction caused the along-shore current to completely reverse and become northerly on the ocean side, while the eddies inside the inlet disrupted and the flow through inlet became complete outflow. The duration of the changes were dependent on the duration of the storm that affected the region, ranging from 3-4 days for Charley and Fay, and up to 12 days for Frances. The momentum balance within the system also changed significantly during storms. Wind stress generated by strong winds was mostly balanced by a combination of barotropic pressure gradient and bottom friction. Baroclinic pressure gradient became important coinciding with increased precipitation or river discharge events in the system and it was more significant in 2004 compared to the 2008 simulations due to the higher precipitation and river discharge (3 times higher). Salinity inside the estuary was significantly affected by the increased precipitation and river discharge especially during the storms. When the residual flow rate at the inlets was directed towards the ocean, the size of low salinity area was at its maximum. Although the salinity within the system was highly affected by precipitation and river discharge, the stratification within the system was generally very weak due to the shallow and dynamic nature of the system except the regions closer to the rivers and creeks. The model results show two potential regions with a very slow flushing rate.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Bilge Tutak.
Thesis: Thesis (Ph.D.)--University of Florida, 2010.
Local: Adviser: Sheng, Y. P.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2012-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2010
System ID: UFE0042341:00001


This item has the following downloads:


Full Text

PAGE 1

1 EFFECTS OF TROPICAL CYCLONES ON CIRCULATION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUARY AND INLET By BILGE TUTAK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010

PAGE 2

2 2010 Bilge Tutak

PAGE 3

3 To my beloved and extremely missed son, Erdem and wife Fatma

PAGE 4

4 ACKNOWLEDG E MENTS My most sincere thanks go to my wife, Fatma and to my parents for their endless love and support. I wish to express my gratitude and sincere appreciation to my advisor Dr. Y. Peter Sheng for the financial and moral support in addition to his excellent guidance through my PhD study. I would like to thank to the members of my supervisory committee Dr. Arnoldo ValleLevinson, Dr. Alex Sheremet, Dr. Edward Phlips and Dr. Rick Gleeson. This research was conducted in the GuanaTolomatoMatanzas National Estuarine Research Reserve System under an award fro m the Estuarine Reserves Division, Office of Ocean and Coastal Resource Management, National Ocean Service, National Oceanic and Atmospheric Administration (NA07NOS4200012). The field work conducted during Fay and Hanna was partially supported by the Flori da Sea Grant Project R/CS 49 and Florida Department of Environmental Protection (Contract No. RM078) via Florida Coastal Ocean Observing System (FLCOOS) Consortium in 2008. I would like to thank Guana Tolomato Matanzas Estuarine Research Reserve for shari ng their resources and many thanks to Kathy Petrinec and Nikki Dix for helping during the field trips and driving the boat. Many special thanks to Drs. Justin R. Davis, Vladimir Paramygin, Yangfeng Zhang, Taeyun Kim for their help especially in early years of my PhD. Thanks to Drew, Andy, Jung Woo, Tianyi, Amy, Emre, Ilgar, Berkay, Cihan, and all the other coastal students for the fun moments they created making this process more bearable.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGEM ENTS ............................................................................................... 4 LIST OF TABLES ............................................................................................................ 7 LIST OF FIGURES .......................................................................................................... 8 LIST OF ABBREVIATIONS ........................................................................................... 11 ABSTRACT ................................................................................................................... 12 1 INTRODUCTION .................................................................................................... 14 2 SIMULATING THE HYDRODYNAMIC RESPONSE OF NCB ESTUARY TO 2004 AND 2008 TROPICAL CYCLONES ............................................................... 17 Abstract ................................................................................................................... 17 Introduction ............................................................................................................. 18 Study Area and Methods ........................................................................................ 19 Model Simulations .................................................................................................. 21 2004 Simulations .............................................................................................. 21 2008 Simulations .............................................................................................. 22 Initial and Boundary Conditions for 2004 and 2008 Simulations ...................... 23 Model Verification ............................................................................................. 23 Change in Salinity and Stratification ....................................................................... 25 Stratification inside the Estuary ............................................................................... 27 Conclusions ............................................................................................................ 29 3 EFFECTS OF HURRICANES ON RESIDUAL CIRCULATION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUARY AND INLET IN 2004 ... 44 Abstract ................................................................................................................... 44 Introduction ............................................................................................................. 45 Study Area and Methods ........................................................................................ 47 Validation of Simulated Residual Flow .................................................................... 50 Spatial Patterns of Residual Circulation .................................................................. 50 During One Typical Tidal Cycle ........................................................................ 50 Changes to the Residual Flow Pattern during Hurricane Charley .................... 51 Changes to the Residual Flow Pattern during Hurricane Frances .................... 52 Changes to the Residual Flow Pattern during Tropical Depression Ivan .......... 53 Changes to the Residual Flow P attern during Hurricane Jeanne ..................... 53 Spatial Patterns of Momentum Balance .................................................................. 54 During One Typical Tidal Cycle ........................................................................ 54 Momentum Balances during Hurricane Charley ............................................... 55 Momentum Balances during Hurricane Frances .............................................. 55

PAGE 6

6 Momentum Balances during Tropical Depression Ivan .................................... 56 Momentum Balances during Hurricane Jeanne ................................................ 57 Time Variation of Momentum Balances at a Station Inside the Inlet ....................... 58 River Discharge and Precipitation ........................................................................... 58 Vorticity and Evolution of Residual Eddies .............................................................. 59 Conclusions ............................................................................................................ 60 4 EFFECT OF TROPICAL CYCLONES ON RESIDUAL CIRCULATION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUARY AND INLET OBSERVATION AND SIMULATION ....................................................................... 77 Abstract ................................................................................................................... 77 Introduction ............................................................................................................. 78 This Study ............................................................................................................... 80 Study Area and Methods ........................................................................................ 82 Model Verification with ADCP observations ............................................................ 84 Spatial Patterns of Residual Circulation .................................................................. 84 During One Typical Tidal Cycle ........................................................................ 84 Changes to the Residual Flow Pattern during Tropical Storm Fay ................... 85 Changes to the Residual Flow Pattern during Tropical Storm Hanna .............. 87 Spatial Patterns of Momentum Balance .................................................................. 88 During One Typical Tidal Cycle ........................................................................ 88 Changes in Momentum Balance during Tropical Storm Fay ............................ 88 Changes in Momentum Balance during Tropical Storm Hanna ........................ 90 Time Variation of Momentum Balance at a Station Inside the Inlet ......................... 90 Conclusions ............................................................................................................ 91 5 CONCLUSION ...................................................................................................... 107 REFERENCES ............................................................................................................ 113 BIOGRAPHICAL SKETCH .......................................................................................... 118

PAGE 7

7 LIST OF TABLES Table page 3 1 Features of four hurricanes affecting the region in 2004. ....................................... 74 3 2 Summary of changes to residual flow structure and their cause ............................ 75 3 3 Summary of changes to residual momentum balance and their cause .................. 76 4 1 Summary of the existing literature ........................................................................ 106

PAGE 8

8 LIST OF FIGURES Figure page 2 1 Study area, model domain and tropical cyclones ................................................... 32 2 2 Comparison of observed (.) and simulated ( solid) water level time series during 2004 simulation period ....................................................................................... 33 2 3 Comparison of observed (.) and simulated (solid) water level t ime series during 2008 simulation period ....................................................................................... 34 2 4 Comparison of observed (.) and simulated (solid) salinity time series during 2004 simulation period ....................................................................................... 35 2 5 Comparison of observed (.) and simulated (solid) salinity time series during 2008 simulation period ....................................................................................... 36 2 6 Comparison of observed (.) and simulated (solid) vertically averaged velocity time series during 2008 simulation time period .................................................. 37 2 7 Comparison of observed (.) and simulated (solid) vertical velocity profiles inside the St. Augustine Inlet in the along inlet direction at selected time instances during 2008 simulation time period ..................................................................... 38 2 8 River discharge and precipitation during 2004 and 2008 simulation periods ......... 39 2 9 Time variation of hurricane winds surface area of salinity at given thresholds and residual flow rate through inlets, during 2004 simulation period .................. 40 2 10 Time variation of tropical storm winds surface area of salinity at given thresholds and residual flow rate through inlets, duri ng 2008 simulation period ................................................................................................................. 41 2 11 Time variation of stratification (bottom surface salinity) bottom salinity, and surface salinity during 2004 simulation period along the ICWW ......................... 42 2 12 Time variation of stratification (bottom surface salinity) bottom salinity, and surface salinity during 2008 simulation period along the ICWW ......................... 43 3 1 Study area, model domain and 2004 hurri canes .................................................... 62 3 2 Cross sectional residual circulation obtained in a rotating elongated rectangular basin for moderate friction case ( =0.5) including the rotational forces obtained with analytical model and CH3D simulation. ........................................ 63

PAGE 9

9 3 3 Spatial pattern of 34 hour low pass filtered vertically averaged residual flow pattern around St. Augustine Inlet region during a regular tidal cycle calculated with and without the nonlinear advection terms ................................. 64 3 4 Spatial pattern of 34h low pass filtered vertically averaged residual flow patterns around St. Augustine Inlet region during Hurricane Frances ................ 65 3 5 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region during a regular tidal cycle t he along channel and the crosschannel direction. ...................................................................................... 66 3 6 Spatial pattern of 34h low pass filtered along channel momentum balance around St. Augustine Inlet region during Hurricane Frances .............................. 67 3 7 Spatial pattern of 34h low pass filtered cross channel momentum balance around St. Augustine Inlet region during H urricane Frances .............................. 68 3 8 Nondimensional time series of 34h low pass filtered along channel (solid line) and cross channel (dashed line) momentum flux terms at a selected station inside the inlet for approximately two month period between 08/06/2004 and 09/31/2004 covering four major hurricanes in 2004. ........................................... 69 3 9 River discharge (Pellicer Creek) and precipitation (Domenico Circle) around St. Augustine Inlet during the simulation period ....................................................... 70 3 10 Momentum balance occurring after hurricanes with considerable precipitation in the along channel and crosschannel direction ............................................... 71 3 11 Variation of cross sectional residual flow at the St. Augustine Inlet at three different transects ............................................................................................... 72 3 12 Evolution of residual eddies, streamlines and vorticity at different phases of Hurricane Frances .............................................................................................. 73 4 1 Study area, modeling domain and tropical stroms in 2008 ..................................... 94 4 2 Comparison of observed (.) and simulated (solid) vertically averaged velocity time series during 2008 simulation time period east west and northsouth dire ctions ............................................................................................................ 95 4 3 Spatial pattern of 34h low pass filtered residual circulation pattern around St. Augustine Inlet region during a reg ular tidal cycle. ............................................. 96 4 4 Spatial pattern of 34h low pass filtered vertically averaged residual flow patterns around St. Augustine Inlet region during TS Fay .................................. 97 4 5 Spatial pattern of 34h low pass filtered vertically averaged residual flow pattern around St. Augustine Inlet region during TS Hanna ............................... 98

PAGE 10

10 4 6 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region during a regular tidal cycle in t he along channel an d the cross channel direction ................................................................................ 99 4 7 Spatial pattern of 34h low pass filtered along channel momentum balance around St. Augustine Inlet region during TS Fay .............................................. 100 4 8 Spatial pattern of 34h low pass filtered cross channel momentum balance around St. Augustine Inlet region during TS Fay .............................................. 101 4 9 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region when TS Hanna closest to the Florida Coast (09/05/2008 12:00 UTC) in along channel and crosschannel directions ......... 102 4 10 Nondimensional time series of 34h low pass filtered along channel (solid line) and cross channel (dashed line) momentum flux terms at a selected point inside the inlet for a month period between 08/15/2008 and 09/15/2008 covering TS Fay (day 235) and TS Hanna (day 250). ...................................... 103 4 11 Observed river dis charge at Pellicer Creek (south of St. Augustine Inlet) and cumulative daily rainfall at Domenico Circle station (12 km south of St. Augustine Inlet) ................................................................................................. 104 4 12 Variation of cross sectional residual flow at the St. Augustine Inlet at three different transects ............................................................................................. 105

PAGE 11

11 LIST OF ABBREVIATIO NS ADCP Acoustic Doppler Current Profiler CH3D Curvilinear Hydrodynamics in 3D GTMNERR GuanaTolomatoMatanzas National Estuarine Research Reserve H Hurricane HRD Hurricane Research Division NCB Northern Coastal Basin NOAA National Oceanic and Atmospheric Administration RMSE Root Mean Square Error SAI St. Augustine Inlet TS Tropical Storm

PAGE 12

12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EFFECTS OF TROPICAL CYCLONES ON CIRCULAT ION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUARY AND INL ET Bilge Tutak December 2010 Chair: Y. Peter Sheng Major: Coastal and Oceanographic Engineering Estuaries along the coast of U.S. are subjected to extreme events such as tropical cyclones, n or easters, and heavy rainfall. These extreme events can bring very drastic and sudden changes to the estuarine environment and habitat. This study focuses on the hydrodynamics of the subtropical Northern Coastal Basin (NCB) estuary during regular (nonstorm) time periods as well as tropical cyclones. Hydrodynamics (water level, currents and salinity) of the system is simulated with a threedimensional model, CH3D (Curvilinear grid Hydrodynamics in 3D), dur ing four major hurricanes (Charley, Frances, Ivan, Jeanne) in 2004 and two tropical storms (Fay and Hanna) in 2008. After successful calibration and verification of using the observed water level, salinity and current (2008) data throughout the estuary, th e model is used to examine the residual circulation, momentum balance, and stratification in order to understand the effects of tropical cyclones on the estuarine hydrodynamics. Model results during a typical tidal cycle show formation of four residual eddies around the headlands of the St. Augustine Inlet due to the nonlinear advection. The dynamic structure of the residual eddies strongly influences the cross sectional residual flow inside the inlet. Analysis of the residual momentum balance simulated by the model reveals that during a typical tidal cycle with weak external forcing, the balance is between the barotropic pressure

PAGE 13

13 gradient and the nonlinear advection. Prior to the landfall of storms, strong northeasterly winds disrupted the ocean side eddies and created a uniform along shore southerly current which forced complete inflow through the inlet due to Ekman transport, but eddies inside the inlet persisted. After the landfall, the wind direction changed to easterly followed by southeasterly winds. T he change in wind direction caused the along shore current to completely reverse and become northerly on the ocean side, while the eddies inside the inlet disrupted and the flow through inlet became complete outflow. The duration of the changes were dependent on the duration of the storm that affected the region, ranging from 34 days for Charley and Fay, and up to 12 days for Frances. The momentum balance within the system also changed significantly during storms. Wind stress generated by strong winds was mostly balanced by a combination of barotropic pressure gradient and bottom friction. Baroclinic pressure gradient became important coinciding with increased precipitation or river discharge events in the system and it was more significant in 2004 compared to the 2008 simulations due to the higher precipitation and river discharge (3 times higher). Salinity inside the estuary was significantly affected by the increased precipitation and river discharge especially during the storms. When the residual flow rate at the inlets was directed towards the ocean, the size of low salinity area was at its maximum. Although the salinity within the system was highly affected by precipitation and river discharge, the stratification within the system was generally very weak due to the shallow and dynamic nature of the system except the regions closer to the rivers and creeks. The model results show two potential regions with a very slow flushing rate.

PAGE 14

14 CHAPTER 1 INTRODUCTION Estuaries form the connection between the high salinity ocean water and fresh riverine water. Estuarine systems are semi closed embayments formed through different morphological processes. Although there are different types of estuaries, they all create living environments for a variety of plants and animal communities. Because they contain the ocean and riverine waters, estuaries generally provide a highly productive water body and have significant economic and recreational values. E stuaries maintain a hydrodynamic balance within the system. The balance is a result of different forces interacting on the system. Tides, rivers, winds waves, and atmospheric events will change the estuarine hydrodynam ics. The circulation is generally forced by the ocean tides and modified by other external forces like wind and density gradients The circulation patterns inside the estuaries determine the transport mechanisms for salt, nutrients and sediments. These pat terns can change over a variety of time scales, from hourly, daily, weekly, monthly, to seasonally, depending on the variation in the forcing functions. It is obvious that extreme events can exert disruptive forces on the estuarine hydrodynamics and water quality, creating possible environmental problems. Environmental problems within the estuary might be exacerbated by such factors as a large urban area within the watershed, increased human intervention, and increased number of point and nonpoint sources These problems might be localized for confined areas, or might affect the water body throughout the estuary, which is directly affected by the circulation characteristics of the estuarine system.

PAGE 15

15 Extreme events like tropi cal cyclones and heavy rainfall can disrupt these circulation patterns and modify the hydrodynamic balance inside the estuary. For example, high precipitation along with heavy river discharge might lead to sharp reductions in estuarine salinity S trong storm winds might completely chang e the flow patterns within tidal inlets. A fter these extreme events it will take some time for the estuarine system to recover to its prestorm conditions, and the recovery time will depend on how quickly the ocean water can be circulate d in to and throughout the system. Therefore understanding the response of the estuarine system to extreme events should help developing better management and mitigation strategies for the estuarine system. This study provides an understanding of the hydrodynamic response of the estuarine water bodies of Northern Coastal Basin (NCB) and Guana TolomatoMatanzas National Estuarine Research Reserve (GTMNERR) to extreme events in 2004 and 2008 by using threedimensional numerical model simulations and observations Chapter 2 investigates the application of the numerical model to the NCB estuary during four hurricanes (Charley, Frances, Ivan, Jeanne) in 2004 and two tropical storms (Fay and Hanna) in 2008. Verification of the 3 D model was done by comparing the simu lated hydrodynamic parameters (water level, current, salinity) to the observed ones at available stations. Variation of salinity and stratification inside the estuary during the hurricanes and tropical storms in 2004 and 2008 was calculated. Residual circ ulation and momentum balance during the hurricanes and tropical storms in 2004 and 2008 inside St. Augustine Inlet, the largest and most active inlet of the NCB, was studied respectively in C hapter 3 and C hapter 4. The variation of the

PAGE 16

16 residual circulation around the inlet during the simulation period was compared to the typical tidal residual circulation The balance between the dominant forces during a typical tidal cycle and the change in the momentum balance due to the strong winds, increased precipitat ion and river discharge during tropical cyclones was studied. A general summary of the results will be discussed in the last chapter.

PAGE 17

17 CHAPTER 2 SIMULATING THE HYDRO DYNAMIC RESPONSE OF NCB ESTUARY TO 2004 AN D 2008 TROPICAL CYCLON ES1 Abstract Northern Coastal Basin (NCB) along the northeast Florida coast was subjected to four major hurricanes (Charley, France, Ivan, and Jeanne) in 2004 and two tropical storms (Fay and Hanna) in 2008. Some of these storms passed through, while the others affected the region despite being far away. This study investigates the hydrodynamic response including the salinity dynamics inside the NCB estuary during 2004 and 2008 tropical cyclones. A threedimensional model, CH3D (C urvilinear grid Hydrodynamics in 3D), was used to simulate the estuarine circulation in the NCB and GuanaTolomatoMatanzas Estuarine Research Reserve (GTMNERR). The model was found to successfully reproduce the observed water level, current, and salinity inside the estuary. Salinity inside the estuary reduced to zero at several stations (Pellicer Creek and Ormond Beach) for extended time period (~15 days) primarily due to the increased freshwater discharge, precipitation and altered residual flow patterns at the three inlets during the hurricanes. Salinity transport patterns inside the estuary show southward residual circulation especially to the south of the St. Augustine Inlet. High salinity ocean water entering the system through the inlets during nonst orm time periods, and the freshwater discharged into the system soon after H Charley and Frances transported south. Recovery time for salinity to return to the prestorm condition was quite long (~15 days) especially for two parts of the system: Pine Island at the north and Halifax River at the south. Despite increased river discharge and precipitation, the stratification 1 To be submitted for publication as Tutak, B. and Y. P. Sheng: Simulating the hydrodynamic response of NCB estuary to 2004 and 2008 tropical cyclones.

PAGE 18

18 throughout the system was never significantly high due to the shallow and dynamic nature of the system. Introduction Tropical cyclones can significantly affect the hydrodynamics, sediment transport (Walker 2001; Lynch 2004), habitat (Boesch 1976; Owen & White 2005), and water quality (Mallin et al. 1999; Paerl et al. 2001) of estuarine systems. For example, Hurricane Dora (1964) breached the barrier island between Matanzas Inlet and Intracoastal Waterway in the area (Powell 2006). Valiela et al. (1998) found that Hurricane Bob destratified the entire Waquoit Bay (MA) and the bay recovered quickly due to rapid flushing rates. Increased precipitation during Hurricane Agnes pushed the higher salinity estuarine water more than 30 miles down the Chesapeake Bay (Sherman 2003). Many twoor threedimensional numerical models have been developed for accurate simulation of the storm surge and inundati on during tropical cyclones (Jelesnianski et al. 1992 ; Westerink et al. 2008; Sheng et al. 2005; Sheng et al. 2010a). However, most of these modeling studies focused on the storm surge simulation over 27 days and did not include the baroclinic effects. Si nce increased river discharge and precipitation during tropical cyclones can lead to significant salinity variations within a system, it is important to include the baroclinic effects in simulating hydrodynamics during tropical storms. Li et al. (2006) si mulated (with 1km grid) the circulation in Chesapeake Bay during Hurricane Isabel and found that high salinity ocean water was forced into the estuary due to strong winds which created very strong vertical mixing, transforming the partially mixed Chesapeak e Bay to a well mixed one. Similarly, response of the York River estuary to Hurricane Isabel was analyzed by Gong et al.

PAGE 19

19 (2007) using a threedimensional hydrodynamic model with a relatively fine grid (250m) over the course of a four month period. Gong et al. found that the system became highly stratified due to the combined effect of high salt influx, and increased freshwater discharge. Model results showed that the system recovered to its pre storm conditions in almost four months. Recently, baroclinic pr ocesses have been incorporated in to the storm surge forecasting system CH3D SSMS (Curvilinear grid Hydrodynamics in 3D Storm Surge Modeling System) for forecasting salinity and storm surge during Tropical Storm Fay in 2008 (Sheng and Paramygin 2010). The main purpose of this study is to use the CH3D SSMS to investigate the hydrodynamic (water level, currents, and salinity) response of the GTMNERR and NCB estuarine systems during the tropical cyclones in two active hurricane seasons in 2004 and 2008. Nu merical results and observed data are compared and used to understand the changes in the estuarine hydrodynamics. Study Area and Methods The modeling study includes almost the entire Northern Coastal Basin (NCB) ( Fi gure 2 1 a), an estuary located on the northeast coast of Florida. The system extends from south of Ponce de Leon Inlet at the south to the east of Jacksonville at the north. The estuarine water body consists of Tolomato River at the north, Matanzas River at the center, and Halifax River at the south along the Intracoastal Waterway (ICWW). ICWW is a very shallow and narrow channel (mean depth=2.7m) inside the estuary. The estuarine water body is connected to the Atlantic Ocean through three tidal inlets: St. Augustine Inlet (tidal range=1.6m) at the north, Matanzas Inlet (tidal range=1.5m) at the center and Ponce de Leon Inlet (tidal range=1.2m) at the south end of the NCB. The depth inside the n avigation channel is up to 15 m inside the St. Augustine Inlet, 12m

PAGE 20

20 inside the Ponce de Leon Inlet, but the depth is only 3 m inside the Matanzas Inlet which is a natural and unmanaged inlet. The NCB system is fed by many small and midsized creeks with relatively small discharges throughout the ICWW. GTMNERR, located inside the NCB, is comprised of a north section (Guana Aquatic Preserve), located to the north of the St. Augustine Inlet and a south section, located between the south of St. Augustine Inlet and south of Marineland. During 2004 and 2008 the area was affected by several different hurricanes and tropical storms ( Figure 2 1 cd). To simulate and understand the change in estuarine hydrodynamics, the Curvilinear Grid Hydrodynamics in 3D (CH3D) model ( Sheng 1987, 1990; Sheng et al. 2010a) is used. CH3D is a robust hydrodynamic model that can be used to simulate the estuarine and coastal circulati on driven by tides, winds, waves, and density gradients. CH3D can accurately resolve the complex geometry and bathymetry of the estuarine system, using a nonorthogonal boundary fitted curvilinear grid in horizontal directions and terrain following sigma g rid in the vertical direction. In CH3D, a secondorder turbulence closure model (Sheng and Villaret 1989) is used to resolve the vertical turbulent mixing, while a Smagorinski type model is used to parameterize the horizontal turbulent mixing. The CH3D mo del has been successfully verified with data from many estuaries throughout the coastal waters of U.S. and Florida, including GTMNERR (Sheng et al. 2008), Chesapeake Bay (Sheng et al. 2010a), Indian River Lagoon (Sheng and Kim 2009), Charlotte Harbor (Kim et al. 2010), and Pensacola Bay (Sheng et al. 2010b). Detailed equations of motion and boundary conditions for CH3D in curvilinear grid can be found in Sheng et al. 2010.

PAGE 21

21 Model Simulations The region was affected by four different storms during 2004 (Char ley, France Ivan, and Jeanne) an d two during 2008 (Fay and Hanna) M odel simulations for both years covering all six storms are conducted. Model simulations were conducted using a high resolution boundary fitted curvilinear grid ( Figure 2 1 b) which extends 150 km north from the south of Ponce de Leon Inlet and 30 km offshore. The minimum grid spacing for the grid is approximately 15 m inside the ICWW to accurately resolve the channel. Total number of cells for the grid was 37250 water, and 22640 floodable cells. Two simulate the three dimensional flow field and resolve the sharp bathymetric changes, 8 sigma layers are used in vert ical. Time step used for all the simulations was 10 s. In order to account for the inundation due to tides and storm surge, flooding anddrying feature of the model was used. 2004 Simulations The region was affected by four major hurricanes in 2004. Hurric ane Charley first made landfall on the west coast of Florida near Ponta Gorda, as a category 4 hurricane, on August 13, 2004. After crossing the state of Florida, Charley emerged from Daytona Beach to the Atlantic Ocean on August 14, 2004. Hurricane Frances made landfall on September 5, 2004 near Hutchinson Island along the southeastern coast of Florida a category 2 hurricane. Less than 2 weeks later, tropical depression Ivan, remnants of Hurricane Ivan which made its first landfall in southern Alabama, emerged to the Atlantic Ocean and made its second landfall on September 21, 2004 at a location very near where Frances made landfall. On September 26, 2004, right after tropical depression Ivan passed through the region, Hurricane Jeanne hit the southeastern coast of Florida.

PAGE 22

22 To include all four hurricanes in this study, we select a twomonth simulation period between August 1, 2004 and October 1, 2004. In the simulations, tides and offshore surge were included as offshore boundary conditions, while the river discharge was used as the upstream river boundary condition. Wind, atmospheric pressure and precipitation were used as model input over the entire domain. Model simulations included the simulation of salinity and baroclinic terms since the precipitation an d river discharge increased quite significantly during the two month period. In addition, to have a stable salinity and density field, a spinup period from the beginning of June to the beginning of August was used. Spinup period was simulated similar to the main simulation time period and included all the forcing functions used for main simulation period. 2008 Simulations In 2008, the area was affected by two tropical storms during a two week period. On August 19, 2008 Tropical Storm Fay made landfall alo ng the southwest coast of Florida. It then crossed the state of Florida in less than 24 hours and emerged to the Atlantic Ocean around Melbourne on August 20, 2008. After hanging around the region with a very slow forward moving speed (57 km/h), TS Fay mo ved northwest and made landfall on late August 21, 2008, near Flagler Beach within the NCB. Two weeks after TS Fay, TS Hanna was the next storm to affect the region while moving approximately 300 km offshore of the east Florida coast. At its closest point, TS Hanna was 200 km offshore on September 5, 2008. Simulation of hydrodynamics in 2008 includes both tropical storms and covered a onemonth period of from August 15, 2008 to September 15, 2008. Similar to the 2004 simulations tides and offshore surge wer e included as offshore boundary conditions,

PAGE 23

23 while the river discharge was used at the upstream boundary of the rivers. Wind, atmospheric pressure and the precipitation were used over the entire domain as model input. In 2008 precipitation and river dischar ge increased significantly during the tropical storms, but it was less compared to the 2004 simulation period. To account for the changes due to precipitation and river discharge, baroclinic terms were included in the simulations as well. To ensure stabili ty of the salinity and density fields throughout the domain, a spinup time period from July 1, 2008 to August, 15, 2008 was used. Initial and Boundary Conditions for 2004 and 2008 Simulations The model boundary conditions and initial conditions were obtai ned from observed data sets and regional scale models. Tidal harmonic constituents for tidal boundaries were obtained from the ADCIRC tidal database for the East Coast (Mukai et al. 2001). Atmospheric pressure and wind data during nonstorm periods are obt ained from the NOGAPS model output, while those during storm time periods are obtained from a combination of winds from NOAA (National Oceanic and Atmospheric Administration) Hurricane Research Divisions (HRD) 10 m Reanalysis H*Wind (Powell et al. 1998) a nd atmospheric pressure from Holland (Holland 1980) analytical model. Observed daily precipitation data available at several stations are interpolated onto the model grid. Observed river discharge data available from USGS is used for river boundary conditi ons at upstream river locations. Model Verification For model verification, simulated water level and salinity were compared at all the available stations inside the domain. The comparisons were made over the main simulation period of 60 days in 2004 and over 30 days in 2008. In general the observed and simulated values showed a very good agreement. The comparisons of water level

PAGE 24

24 for selected stations in 2004 and 2008 are given in Figure 2 2 and Figure 2 3 respectively. The simulated water level values are mostly within 1520 cm of the observed water level values including during the tropical storms and hurricanes. The average normalized root mean square (RMS) error for all the water level stations for both during 2004 and 2008 are approximately 10%. Similarly, the simulated salinity over 60 days in 2004 and 30 days in 2008 are generally in agreement with the observed salinity values, as shown in Figure 2 4 and Figure 2 5 respectively. The simulated salinity values are generally within 36 ppt of the observed values, while the mean normalized RMS error for all the salinity stations is around 25%. It is important to note that tropical cyclones are generally associated with increased precipitation and river discharge which can cause sudden changes in the local salinity. Hence, it is difficult to accurately simulate the salinity without accurate precipitation and river discharge measurements and surface runoff created by heavy rainfall. In addition to the water level and salinity observation, currents inside the deep navigation channel of St. Augustine Inlet were obtained with an Acoustic Doppler Current Profiler (ADCP) during Tropical Storms Fay and Hanna in 2008. The ADCP ( Figure 2 1 a) was bottom mounted and recorded the velocity profile over the water column at every 0.5 m at a ping rate of 1.5 s, averaging 900 samples per ensemble. The vertically averaged velocities simulated by the model at the same location agree well with the measured ones in both east west and northsouth directions ( Figure 2 6 ). In addition, comparison of several vertical velocity profiles in the dominant flow direction (east west) are given at selected times in Figure 2 7 where model results show better agreement with data during flood and ebb than during slack tides. This is because that,

PAGE 25

25 during flood and ebb tides, the tidal boundary layer is thicker and hence better simulated. The average RMS error for the simulated east west velocities at all 8 vertical layers is approximately 20 cm/s (8% of the maximum current) and less than 10 cm/s (8% of the maximum curr ent) for the northsouth velocities. The average RMS error for vertically averaged velocities in east west direction is less than 20 cm/s and less than 5 cm/s for the north south velocities. Change in Salinity and S tratification During 2004 and 2008 simulation periods, precipitation and river discharge increased substantially during tropical cyclones ( Figure 2 8 ). Increased precipitation and river disch arge throughout the estuary created significant variation in the salinity field. During normal conditions, most part of the subtropical estuary flushes rather quickly (~12 days). However, some localized sections of the system, especially near the midsize d creeks located away from the inlets have longer residence times (~15 days) (Sheng et al. 2008). During tropical cyclones, the structure of the residual flow throughout the estuary as well as the inlets changes completely. During tropical cyclones in 200 4 and 2008, the residual flow around the inlets changed significantly which controlled the distribution and recovery of the estuarine salinity (Tutak and Sheng 2010a, 2010b). In order to understand the variation of salinity within the estuary, we examine t he variation of the total surface area of estuarine water with three distinct salinity ranges: 010 ppt (low salinity), 10 20 ppt (moderate salinity), and 20 ppt and above (high salinity). The areas of these three salinity ranges change during tropical cyc lones due to increased precipitation, river discharge and change in the residual flow structure throughout the estuary. Changes of these three salinity areas in 2004 and 2008 are shown in Figure 2 -

PAGE 26

26 9 b and Figure 2 10b, respectively. These changes in salinity areas are highly dependent on both the river discharge and the precipitation shown in Figure 28 In addition, the sudden changes in the areas seem to be highly correlated with the residual flow rate ( Figure 2 9 c, Figure 2 10c) at each inlet espec ially during the hurricanes ( Figure 2 9 a, Figure 2 10a) and tropical storms. In 2004, prior to Julian day 230, the salinity field appears to be rather stable. Around and after Julian day 230, the estuarine system seems to be varying temporally with semi diurnal tide. A small jump in the low salinity (0 10ppt) area is due to a change in the system caused by Hurricane Charley. After that, salinity in the system was stable until day 245, when Hurricane Frances started to affect the region. Between day 245 and 250, the low salinity area started to reduce, while the high salinity area increased, due to increase in residual inflow (negative flow rate) at the inlets due to the easterly winds. Right between day 249 and 250, Hurricane Frances made landfall and started moving towards north across Florida. Theref ore the barotropic pressure gradient created due to the strong winds relaxed in the system, along with the increase in precipitation and river discharge. The residual flow rate through the inlets reversed and became positive (out of inlet). The amount of r esidual flow rate was quite significant (up to 1000 m3/s), thus created a quicker change in the salinity. The high salinity (20 ppt and above) area reduced approximately to onefifth of its value in one day. At the same time, the low and moderate (10 20 ppt) salinity areas doubled in coverage. The low salinity area stayed high for approximately 15 days until around day 265, when it was almost equal to the high salinity area, similar to the prestorm condition around day 240. Around day 270 during Hurricane Jeanne, the low salinity area first showed a slight

PAGE 27

27 reduction and then jumped back to its prestorm size of 50 km2 coverage. The moderate salinity area did not show significant variation with time. In 2008, the variation of salinity areas was a little di fferent compared to the 2004 season. The high salinity area started to reduce around day 234 due to the effect of increased precipitation and river discharge. The residual flow in St. Augustine Inlet was into the estuary (negative) between day 234 and 235, and out of the estuary between day 235 and 236. However, the other two inlets never showed inflow of water during TS Fay due to the direction of the winds. The salinity did not seem to be affected much by the residual flow during TS Fay, as it occurred wi th Hurricane France in 2004, since Fay was a weaker storm. The low salinity area increased up to 20 km2 until day 240 and remained stable until day 245, when the precipitation and river discharge became negligible. After day 245, the moderate salinity area started to increase due to mixing with high salinity ocean water. After day 250 during TS Hanna, the low salinity area started to decrease with small fluctuations and it reduced down to 10km2 until the end of the simulation period. In the mean time, the high salinity area showed a small decrease between day 251 and 253, but then recovered up to 60 km2. At the end of the 30 day simulation, the moderate salinity area reached 30km2, which was formed due to the mixing of high and low salinity water inside the estuary. Stratification inside the E stuary The relatively small size of the estuary combined with the tidally dominated flow leads to a generally well mixed salinity and density field throughout the system with the exceptions around the rivers and creeks. However, since the NCB estuary is subtropic al, amount of freshwater is generally limited. To understand and determine how stratification was affected during tropical cyclones inside the NCB, the difference

PAGE 28

28 between the bottom and the surface salinity throughout the modeling domain was calculated. To be representative, the variation of stratification throughout the ICWW is given in Figure 2 11a for 2004, and Figure 2 12a for 2008 simulations. The difference between the bottom and surface salinity was generally less than 1 ppt throughout the estuary. The biggest differences (dark patches) were fou nd around the river mouths. The most stratified region was at the mouth of Tomoka River (~40 km from Ponce de Leon Inlet) which had one of the strongest freshwater discharge values within the system, and was the closest river to the ICWW as well. During t ropical cyclones, the stratification disappeared even around the rivers and creeks. This was especially obvious during Hurricane Frances (~ day 250) in 2004 and TS Hanna (~ day 250) in 2008 due to the strong vertical mixing caused by the strong wind throug hout the estuary. In addition, the bottom ( Figure 2 11 b and Figure 2 12 b) and surface ( Figure 2 11c and Figure 2 12c) salinity with respect to distance and time inside the ICWW are plotted to understand the transport of the high and low salinity water inside the estuary. Model results showed that the freshwater especially from the Tomoka River and Pellicer Creek (~75 km from Ponce de Leon Inlet) moved towards south inside the ICWW during the simulation period in 2004. During the 2008 simulation period, the freshwater from Pellicer Creek did not reach the ICWW, but the freshwater from Tomoka River was transported south towards Halifax River. The transport was more prominent after the tropical cyclones, between day 228 and 240 after Hurricane Charley, between day 250 and 265 after Hurricane Frances in 2004, and between 236 and 250 after TS Fay in 20 08. Similarly, the high salinity ocean water that entered the estuary from St. Augustine Inlet (~100 km) and Matanzas I nlet (~80 km) moved mostly

PAGE 29

29 towards the south in both simulation periods due to the higher mean water level at the north part of the syste m The sudden increase of high salinity water around day 249 during the landfall of Hurricane Frances in 2004 coincided with t he strong northeasterly winds and the negative (inflow) residual flow. In the ensuing day 250, the high salinity water inside the ICWW decreased suddenly due to the relaxation of wind setup, and the completely reversed (positive) residual flow through the inlets. At the north section of the estuarine system (north of St. Augustine Inlet), the variation of salinity was controlled mos tly by the small amount of freshwater discharge from several creeks. The high salinity ocean water did not reach up the estuary beyond 1020 km north of the St. Augustine Inlet. In addition, Figure 2 11 and Figure 2 12 along with Figure 2 9 and Figure 2 10 showed that once the system received the low salinity freshwater dis charge, the flushing time of the system, especially at the southern (around Halifax River) and northern (near Pine Island) sections of the estuary, was considerably longer (more than 2 weeks). The slow flushing was observed during both 2004 and 2008 simulation periods, but it was more obvious during 2004 due to the higher freshwater (3 times higher during maximum river discharge) input to the system compared to the 2008 simulation period. The same two sections of the system were shown by Sheng et al. (2008) to be the slowest flushing sections of the estuarine system. Conclusions Dynamics of water level, salinity and stratification throughout the GTMNERR and NCB have been investigated during 2004 and 2008. The system was affected by four major hurricanes (Charley, France, Ivan, Jeanne) in 2004 and two tropical storms in 2008 (Fay and Hanna). A threedimensional baroclinic and flooding anddrying

PAGE 30

30 numerical model CH3D was used to simulate the estuarine hydrodynamics of water level, salinity, and flow velocities. The simulated water level, salinity and flow velocities are found to agree with the observed ones quite well. The salinity inside the system was subjected to dramatic reductions especially during the storms. Some local areas of the system such as Pellicer Creek and Halifax River were reduced to zero salinity which persisted for more than 15 days. The increased precipitation and river discharge flooded the system with low salinity freshwater. In addition, the residual flow at the three inlets was significan tly modified by the tropical cyclones and prevented the high salinity ocean water from entering into the system. The variation of salinity along the ICWW showed that at the south of St. Augustine Inlet, the salinity transport was towards the south due to the higher mean water level at the north of the domain. This was observed at several different instances throughout the simulation period: the transport of high salinity ocean water entering the domain from inlets in to the system right before H Charley and Frances, and the transport of low salinity freshwater discharged from the rivers right after H Charley and Frances. Two areas of the NCB affected by the increased freshwater discharge and precipitation showed long recovery times (~15 days) due to their di stance to the inlets and the modified circulation patterns inside the inlets. The northern section of the system around Pine Island (SWMP) and the southern section of the system near Halifax River showed persistent low salinity conditions for more than 2 w eeks. Similarly, the flushing time for these two regions was shown by Sheng et al. (2008) to be around 15 days and designated as the two slowest flushing segments of the estuarine system. Stratification inside the system was negligible except at river mout hs, although the salinity of the system went through dramatic changes.

PAGE 31

31 Stratification was reduced during tropical cyclones due to increased vertical mixing caused by strong winds.

PAGE 32

32 a) b) c) d) Figure 2 1 Study area, (a) map showing the study area of Northern Coastal Basin and GTMNERR. Red star at the St. Augustine Inlet denotes the ADCP location used for current measurements; (b) CH3D modeling domain on the east coast of Florida; (c) hurricanes in 2004 ; (d) tropical s torms in 2008

PAGE 33

33 a) b) c) d) Figure 22 Comparison of observed (.) and simulated (solid) water level time series during 2004 simulation period at a) Port Orange; b) Ormond Beach; c) Bings Landing; d) Vilano Beach J u l i a n D a yW a t e r L e v e l ( c m ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yW a t e r L e v e l ( c m ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yW a t e r L e v e l ( c m ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yW a t e r L e v e l ( c m ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d

PAGE 34

34 a) b) c) Figure 23 Comparison of observed (.) and simulated (solid) water level time series during 2008 simulation period at a) Ponce de Leon Inlet; b) San Sebastian (SWMP); c) Pine Island (SWMP) J u l i a n D a yW a t e r L e v e l ( c m ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yW a t e r L e v e l ( c m ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yW a t e r L e v e l ( c m ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d

PAGE 35

35 a) b) c) d) Figure 24 Comparison of observed (.) and simulated (solid) salinity time series during 2004 simulation period at a) Fort Matanzas (SWMP); b) Pellicer Creek (SWMP); c) Crescent Beach; d) San Sebastian (SWMP) J u l i a n D a yS a l i n i t y ( p p t ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d J u l i a n D a yS a l i n i t y ( p p t ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d J u l i a n D a yS a l i n i t y ( p p t ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d J u l i a n D a yS a l i n i t y ( p p t ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d

PAGE 36

36 a) b) c) Figure 25 Comparison of observed (.) and simulated (solid) salinity time series during 2008 simulation period at a) Pellicer Creek (SWMP); b) San Sebastian (SWMP); c) Pine Island (SWMP) J u l i a n D a yS a l i n i t y ( p p t ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d J u l i a n D a yS a l i n i t y ( p p t ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d J u l i a n D a yS a l i n i t y ( p p t ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 0 1 0 2 0 3 0 4 0 O b s e r v e d S i m u l a t e d

PAGE 37

37 a) b) Figure 26 Comparison of observed (.) and simulated (solid) vertically averaged velocity time series during 2008 simulation time period in a) east west; b) northsouth directions J u l i a n D a yV e l o c i t y ( c m / s ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yV e l o c i t y ( c m / s ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 6 0 3 0 0 3 0 6 0 O b s e r v e d S i m u l a t e d

PAGE 38

38 Figure 27 Comparison of observed (.) and simulated (solid) vertical velocity profiles inside the St. Augustine Inlet in the along inlet direction at selected time instances during 2008 simulation time period

PAGE 39

39 a) b) Figure 28 River discharge and prec ipitation during a) 2004 simulation period with noticeable increase during Hurricane Charley (day 225230), Hurricane Frances (day 248260), Hurricane Ivan (day 263268), Hurricane Jeanne (day 268275); b) 2008 simulation period with noticeable increase during Tropical Storm Fay (day 233248), Tropical Storm Hanna (day 248255)

PAGE 40

40 a) b) c) Figure 29 Time variation of a) hurricane winds; b) surface area of salinity at given thresholds; c) residual flow rate through inlets, during 2004 simulation period 1 0 ( m / s ) J e a n n e I v a n F r a n c e s C h a r l e y J u l i a n D a y S u r f a c e A r e a ( k m 2 ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 0 2 0 4 0 6 0 8 0 1 0 0 0 < S < 1 0 1 0 < S < 2 0 2 0 < S J u l i a n D a y F l o w R a t e ( m 3 / s ) 2 2 0 2 3 0 2 4 0 2 5 0 2 6 0 2 7 0 1 5 0 0 1 0 0 0 5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 S t A u g u s t i n e I n l e t M a t a n za s I n l e t P o n c e d e L e o n I n l e t

PAGE 41

41 a) b) c) Figure 210. Time variation of a) tropical storm winds; b) surface area of salinity at given thresholds; c) residual flow rate through inlets, during 2008 simulation period 1 0 ( m / s ) H a n n a F a y J u l i a n D a y S u r f a c e A r e a ( k m 2 ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 0 2 0 4 0 6 0 8 0 1 0 0 0 < S < 1 0 1 0 < S < 2 0 2 0 < S J u l i a n D a y F l o w R a t e ( m 3 / s ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 1 5 0 0 1 0 0 0 5 0 0 0 5 0 0 1 0 0 0 1 5 0 0 S t A u g u s t i n e I n l e t M a t a n za s I n l e t P o n c e d e L e o n I n l e t

PAGE 42

42 a) b) c) Figure 211. Time variation of a) stratification (bottom surface salinity); b) bottom salinity c) surface salinity during 2004 simulation period along the ICWW

PAGE 43

43 a) b) c) Figure 212. Time variation of a) stratification (bottom surface salinity); b) bottom salinity c) surface salinity during 2008 simulation period along the ICWW

PAGE 44

44 CHAPTER 3 EFFECTS OF HURRICANE S ON RESIDUAL CIRCUL ATION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUARY AND INL ET IN 20042 Abstract In 2004, four hurricanes (Charley, Frances, Ivan, and Jeanne) affected the subtropical Northern Coastal Basin (NCB) estuary and St. Augustine Inlet (SAI). This study investigates the changes in residual circulation and momentum balance inside the NCB and SAI during these hurricanes using a threedimensional numerical model CH3D. Residual circulation simulated by CH3D is verified with results of a 3D analytical model. During typical tidal cycles with weak external forces, simulated residual circulation shows two counter rotating (~1520 cm/s) eddies on the estuary side and two (~5 10 cm/s) eddies on the ocean side of the SAI. This structure results in cross sectional residual flow to change along the inlet. During hurricanes, the residual circulation changed significantly the strong northeasterly winds disrupted the oceanside eddies and created a southerly along shore current, while the ensuing easterly winds created an inflow throughout the mouth of the inlet followed by an outflow throughout the inlet and a northerly along shore flow due to southeasterly winds. The time for flow recovery was very short for Hurricanes Charley and Ivan (~12 days), but much longer (~10 days) for Hurricanes Frances and Jeanne. Comparison of the terms in the horizontal momentum equations show that over a typical tidal cycle the balance is predominantly between barotropic pressure gradient and nonlinear advection. Wind stress during hurricanes became significant part of the balance, causing the eddies to disrupt. Bottom friction and baroclinic pressure gradient increase significantly to 2 To be submitted as Tutak, B. and Y. P. Sheng: Effect of Hurricanes on Residual Circulation and Momentum Balance in a Subtropical Estuary and Inlet in 2004

PAGE 45

45 contribute to the balance. Comparing residual circulati on calculated with and without the nonlinear advection shows that the residual circulation is caused by nonlinear advection. Introduction Residual flow in an estuarine system may be created due to a variety of forces such as wind, river discharge, nonlinearities, etc., and is significantly influenced by the physical characteristics of the estuarine system. Fischer et al. (1979) defined the residual flow to be the average flow over a tidal cycle, but cautioned that each tidal cycle is different from another one. Kjerfve (1978) used observed data obtained from North Inlet in North Carolina, a well mixed and shallow estuary with two deep channels separated by a shoal in between, to examine the relationship between residual flow and variations in bathymetry. Kj erfve found a pattern of outflow in the deep channel and inflow over the shallow region, opposite to the pattern proposed by Fischer (1976) for a partially mixed, density driven estuary with laterally varying bathymetry. Li and ODonnell (1997), using a tw o dimensional analytical model in a channel with laterally varying (v shaped) bathymetry, obtained similar residual flow pattern as that obtained by Kjerfve (1978). They found that the residual flow was the result of the nonlinearities created by inertial effects and tides. Li and ODonnell (2005) used a similar analytical model to study the effect of channel length (parameterized by channel depth and tidal frequency) on the residual circulation. The long channels, with progressive waves, show a residual flow structure similar to that in a well mixed estuary with laterally varying bathymetry, while the short channels with standing waves show inflow at the deep channel and outflow over the shallow shoals.

PAGE 46

46 The residual flow through the mouth of the St. Augus tine Inlet (SAI), Florida was measured by Webb et al. (2007) whose data support the short channel theory postulated by Li and ODonnell (2005). In addition to the residual flow structure, Webb et al. found that the force balance inside the inlet was domina ntly between barotropic pressure gradient and nonlinear advection. Bottom friction and Coriolis force are also known to modify the tidal residual circulation. Winant (2008) using a threedimensional analytical model inside an elongated rectangular channel showed, w hen rotation is considered, the tidally generated residual velocities yielded larger values compared to the cases where rotation is not considered. The rotational effects become very important at low friction conditions, even when the channel is v ery narrow. At high friction conditions, the residual flow structure mimics found in long channels (ValleLevinson et al. 2003). Tee (1976) examined the tidally generated residual circulation inside the Minas Basin of Bay of Fundy, using a two dimensional vertically integrated numerical model, and found that the residual circulation and the formation of the residual eddies was largely due to the nonlinear advection terms. Hench and Luettich (2003) studied the momentum balance around the shallow Beaufort Inl et in North Carolina using a two dimensional vertically integrated numerical model, and found that the pressure gradient was balanced by the nonlinear advection in the flood phase, while the bottom friction was in balance with nonlinear advection in the ebb phase. Episodic events such as storms, associated with strong winds and increased river discharge, might affect the residual circulation. Liu (1992) used data observed inside the Indian River Lagoon, Florida and a one dimensional numerical model to show that

PAGE 47

47 barotropic pressure gradients created by storms might be on the same order as the tidal pressure gradients and drastically alter the residual exchange patterns. ValleLevinson et al. (2002) showed that the flow structure at the mouth of the Chesape ake Bay was modified from outflow at the deep channel and inflow over the shoals, to a complete outflow throughout the entire mouth during Hurricane Floyd, due to the increased freshwater discharge, increased wind speed and favorable wind direction. The ai m of this study is to obtain a comprehensive understanding on the effects of extreme events such as tropical cyclones on the residual circulation and the momentum balance inside a subtropical estuary (Northern Coastal Basin or NCB) and tidal inlet (St. Aug ustine Inlet or SAI) in Florida, using a threedimensional hydrodynamic model CH3D (Sheng 1987, Sheng et al. 2010) to simulate the August October period in 2004 when four major hurricanes affected the area. Study Area and Methods St. Augustine Inlet ( Figure 31 a) is a subtropical and tidally dominated inlet along the east coast of Florida. It is a managed inlet including a dredged navigation channel to the south of the inlet with depths up to 15 m. The depth towards the north shore of the inlet is much shallower w ith the maximum depth around 5 m. The inlet is dominated by semi diurnal tides with tidal range reaching up to 1.7 m inside the inlet. SAI is directly connected to the Intracoastal Waterway (ICWW) which includes Tolomato River at the north, Matanzas River and Salt Run (an embayment) at the south. The inlet and the connecting branches are all parts of the NCB estuarine system ( Figure 31 b) which extends from east of Jack sonville at the north to the Ponce de Leon Inlet at the south. The system consists of the very shallow and narrow ICWW and two other inlets: Matanzas and Ponce de Leon. GuanaTolomatoMatanzas National Estuarine

PAGE 48

48 Research Reserve (GTMNERR) is part of the NC B estuary, established to support and provide opportunities for public education, research, and enhanced management of the estuarine habitat (FDEP 2009). Five observation stations maintained by GTMNERR provide real time hydrodynamic, water quality, nutrien t and meteorological information. The flow within the NCB is forced by tides mostly from the SAI. Tidal flow through the SAI is generally 4 5 times larger than the other two inlets. During the 2004 hurricane season, four major hurricanes ( Figure 31 c, Table 3 1 ) (Charley, Frances, Ivan, and Jeanne) affected the regi on. To understand the complete spatial and temporal variation of the residual circulation and the momentum balance inside the SAI during these hurricanes, the CH3D ( C urvilinear grid H ydrodynamics in 3D ) model ( Sheng 1987, 1990 ; Sheng et al. 2010) is used f or the hydrodynamic simulations (Tutak and Sheng 2010a). CH3D model uses horizontally boundary fitted curvilinear grid and a vertically terrainfollowing sigma grid to enable accurate resolution of the complex geometry and bathymetry of estuaries. Moreover CH3D contains a robust turbulence closure model (Sheng and Villaret 1989) for vertical turbulent mixing and a Smagorinski type model for horizontal turbulent mixing. CH3D can be used to calculate the circulation driven by tides, winds and density gradients. In addition, the model can allow flooding anddrying (Davis and Sheng 2003) and accept largescale and hurricanescale winds and precipitation to calculate storm surge and coastal inundation. Detailed explanation of equations and boundary conditions can be found in Sheng et al. (2008) and Sheng et al. (2010a). To incorporate the prestorm and post storm changes to the estuarine system, the model simulations were conducted over a two month period between 08/01/2004 and

PAGE 49

49 10/01/2004 with a two month spinup period between 06/01/2004 and 08/01/2004. Although the residual circulation and momentum balance around SAI are of primary interest, the model domain for the hydrodynamic simulations includes almost the entire NCB to ensure accuracy of the simulation. The model uses a high resolution boundary fitted curvilinear grid, with a minimum cell size of 15 m inside the estuary, and allows the flooding and drying process. 37250 water and 22640 floodable cells comprise the model grid ( Figure 31 d ). In order to capture the bathymetry changes from shallow channels inside the estuary to deep offshore and threedimensional velocity structure, four sigma layers are used. The model sim ulations were conducted with 10 s time step. CH3D uses tidal constituents obtained from ADCIRC tidal database for the East Coast (Mukai et al. 2001) along the offshore boundary. The simulations include forcing by the wind and atmospheric pressure as well. During nonstorm periods, wind and atmospheric pressure are obtained from the Navy Operational Global Atmospheric Prediction System (NOGAPS) model output. During hurricanes, the 10 m reanalysis H*Wind (Powell et al. 1998) wind snapshots generated by NOAA ( National Oceanic and Atmospheric Administration) Hurricane Research Division (HRD) are used in combination with the NOGAPS wind (as background wind) (Paramygin, 2009). Atmospheric pressure during storm periods are obtained from the Holland B (Holland 1980) analytical wind model because the H*Wind reanalysis winds do not include the atmospheric pressure. In addition, for simulating the baroclinic circulation, river discharge at the upstream river boundaries and spatially interpolated precipitation based on data obtained at a few stations are used in the model simulation.

PAGE 50

50 Validation of Simulated Residual Flow Before trying to understand the residual circulation in a region with very complex geometry and bathymetry due to the interaction of different forces, nu merically simulated residual circulation obtained from CH3D is compared with the circulation obtained by Winant (2008) using an analytical model for an elongated, rotating rectangular basin with a laterally varying depth. Winant obtained residual circulati on for different cases of friction and rotational forces (no friction, moderate friction, high friction cases with rotation or no rotation). Although all cases are simulated with CH3D, only comparison of moderate friction case with rotation is presented in this paper. As shown in Figure 3 2 the numerical model results show excellent agreement with the analytical model results. This validation proves that the CH3D has robust physics and numerical algorithms for simulating threedimensional residual circulation. Although analytical models can provide insight for understanding ideal and simplified basins, they cannot be applied to real estuaries with com plex bathymetry and geometry, stratification and variable turbulent eddy viscosity. Therefore, CH3D will be used for simulating the residual circulation in NCB, a real and complex estuary. Spatial Patterns of Residual Circulation During One Typical Tidal Cycle Tidal residual circulation calculated inside the SAI using CH3D show formation of four residual eddies ( Figure 3 3 a). Two stronger eddies form i nside the estuary with currents up to 1520 cm/s, while two weaker eddies (with currents ~510 cm/s) form on the ocean side of the inlet. Inside the estuary, a clockwise eddy in the north and a counterclockwise eddy in the south are formed. Conversely, on the ocean side, a clockwise eddy in the south and a counterclockwise eddy in the north are found. Across

PAGE 51

51 the mouth of the inlet, residual currents show inflow through the deep navigation channel and outflow over the shallow north section as was observed by Webb et al. (2007). The model results show that the cross sectional residual flow pattern observed by Webb et al. on a transect at the mouth of the inlet was part of the two threedimensional residual eddies inside the inlet. Therefore the cross sectional residual flow pattern at the mouth of the inlet is prone to changes with the location of the selected transect ( Figure 3 11 ). The residual eddies ins ide the inlet are important because they determine the distribution and transport of the water mass inside the estuary, and control the cross sectional flow structure at the mouth of the inlet. Understanding the structure and dynamics of these eddies is es sential to the understanding of cross sectional residual flows. The formation of these residual eddies has been shown to depend on the geometry and shape of the channel (Fischer et al. 1979), tidal flow interacting with the local bathymetry (Zimmerman 1978), and the nonlinear advection (Tee 1977). Using CH3D model results, we confirm that the formation of the residual eddies around SAI is mainly due to the nonlinear advection, while geometry and bathymetry affected the detailed structure of the eddies. Figure 3 3 b shows that the residual currents are negligible (~12 cm/s) when the nonlinear advection terms in the model are excluded from the calculations Changes to the Residual Flow Pattern during Hurricane Charley Hurricane Charley made landfall in Charlotte Harbor area on August 13, 2004. It moved across Florida and emerged to the Atlantic Ocean around Daytona Beach early morning on August 14, 2004. Si nce Hurricane Charley was a very compact and fast moving storm, changes to the residual flow around the SAI were. When the storm reached the NCB region, the winds were 9 m/s from northeast which quickly changed to

PAGE 52

52 northwest and then southwest. Due to the s trong winds, the weak eddies on the ocean side of the inlet were disrupted into a unidirectional southerly along shore current up to 10 m/s, while the eddies inside the inlet did not change their structure or strength. Hurricane Charley was moving so fast that the strong wind conditions in the region did not persist more than 6 hours. After the winds relaxed the system returned to its prestorm condition in 48 hours. Changes to the Residual Flow Pattern during Hurricane Frances Hurricane Frances made landfall approximately 325 km away from the region on the east coast of Florida, but the strong tropical storm strength winds reached the area due to its large size. The wind started blowing from northeast at approximately 7 m/s fou r days before the landfall and persisted until two days before the landfall. As a result, the residual flow structure around the region started to change four days before the landfall first with the disruption of the weak ocean side eddies into a unidirect ional southerly along shore flow, followed by an increased residual flow into the inlet from the ocean. When the wind started to intensify (up to 12 m/s) from the northeast, the southerly along shore currents reached a maximum of 20 cm/s ( Figure 3 4 a). In addition, the strong northeasterly weakened the north eddy inside the inlet (speed ~12 cm/s), and the residual inflow coming through the inlet moved t owards Matanzas River revolving around the south eddy without disrupting the south eddy. Wind speed reached approximately 14 m/s and the wind direction turned to southeast in two days following the landfall. The system responded quickly to the increasing w ind speed and changing wind direction. The along shore currents on the ocean side completely reversed and became northerly with speeds up to 25 cm/s

PAGE 53

53 ( Figure 3 4 b). Both eddies inside the inlet completely vanished and flow inside the inlet became complete outflow, supported by the flow from Matanzas and Tolomato Rivers. The structure of the residual flow inside the inlet resembles that of a tidal ebb flow. The winds relaxed in the next 24 hours and the system recovered to its prestorm condition almost 10 days after the residual flow structure started to change. Changes to the Residual Flow Pattern during Tropical Depression Ivan H Ivan made landfall in A labama and Florida border (Sheng et al. 2010b), but later reemerged to the Atlantic Ocean around Maryland and returned to the east coast of Florida as a weak storm. During the passage of a much weakened Ivan, the structure of the residual circulation did n ot change significantly due to the diminishing wind. While Ivan was close to the east coast of Florida, the winds were blowing mostly from the northeast with speed up to 6 m/s. The two eddies on the ocean side were disrupted, and the along shore current became southerly as was during H Charley. Changes to the Residual Flow Pattern during Hurricane Jeanne Less than a week after Ivan made its second landfall, when the system was yet to recover to its preIvan conditions, H Jeanne made landfall at almost the same location where H Frances made the landfall. Before H Jeanne made landfall the ocean side of the inlet was still showing southerly along shore flow with speeds up to 15 cm/s while the wind was blowing at 7 m/s from northeast ( Figure 3 4 a). Since Hurricanes Frances and Jeanne had similar track and structure, changes to the residual circulation around the system were very similar. During H Jeannes landfall the residual flow structure changed drastically around the area. The winds turned from northeast to east while the speeds went up to 19 m/s. The residual flow on the ocean side started to reverse a northerly flow due to relaxed

PAGE 54

54 a long shore wind. The inlet mouth at the moment did not show inout flow condition. After the landfall, wind speed in the region increased to almost 24 m/s and the wind direction became southeasterly. This wind enhanced the northerly along shore residual cu rrents that started to form during landfall and increased the magnitude up to 25 cm/s. The residual flow became complete outflow through the mouth of the inlet and the structure of the residual flow inside the inlet was very similar to a tidal ebb flow as depicted (Figure 7.6) by Fischer et al. (1979), where all the flow was directed towards the mouth of the inlet. This structure was very similar to the one during H Frances ( Figure 3 4 b). A short summary of changes to the residual circulation with respect to the other forces during the simulation period is given in Table 32 Spatial Patterns of Momentum Balance During One Typical Tidal Cycle During tidally dominated periods, the residual momentum balances in along channel (~15 counterclockwise from E W) (Equation 1) and cross channel (Equation 2) directions in the region around the inlet are primarily between barotropic pressure gradient and nonlinear advection terms throughout the region, especially around the inlet, which are at least an order of magnitude higher than the other terms ( Figure 3 5 a b). The other significant terms for the momentum balance are bottom friction and baroclinic pressure gradient. Bottom friction is important especially around the mouth of the inlet where the bathymetry is shallow around tidal flats (ebb shoals). On the other hand, along channel baroclinic pressure gradient during the simulation period is significant mostly inside the Tolomato River along the east coast of ICWW. The cross channel baroclinic pressure gradient is strong inside the mouth of the inlet, due to the

PAGE 55

55 exchange between ocean and estuarine waters. The remaining terms are generally negligible. + ( / ) + ( / ) + + + + = 0 Eq. (3 1) + ( / ) + ( / ) + + + + + = 0 Eq. (3 2) Momentum B alances during Hurricane Charley During Hurricane Charley, the momentum balance was comparable to that during a typical tidal cycle, with slight changes in the wind stress and baroclinic pressure gradient, mostly in the cross channel direction due to the i ncreased wind stress. The bottom friction showed a slight increase in cross channel direction around the tips of headlands on the ocean side due to the southerly along shore current created. Momentum Balances during Hurricane Frances The momentum balance s tarted to show slight changes throughout the region two days prior to the landfall until right before the landfall. Right before the landfall, the wind increased up to 12 m/s from northeast. The along channel wind stress inside the inlet and the Tolomato River increased significantly, thus creating a positive momentum flux ( Figure 3 6 a) which is balanced by increased negative barotropic pressure gradient due to the wind set up. Similarly, in the cross channel direction, a positive momentum flux is created mostly inside the inlet due to the wind stress ( Figure 3 7 a), which is balanced by a combination of barotropic pressure gradient and bottom friction. Bottom friction is significant especially on the ocean side of the inlet. In addition to these terms, baroclinic pressure gradient gained some s trength in the along channel and cross channel

PAGE 56

56 directions, due to the increased freshwater flow into the system which modified the residual momentum balance at the mouth of the inlet. After the landfall of Frances, the spatial structure of the wind stress in the along channel direction did not change significantly except a slight increase inside the Tolomato River ( Figure 3 6 b). The along channel wind s tress is balanced by the barotropic pressure gradient. Bottom friction reversed due to reversal of the residual flow at the mouth of the inlet. Because the wind changed direction from northeast to southeast, the along channel wind stress became completely negative throughout the region ( Figure 3 7 b). The strong wind stress was mainly balanced by the bottom friction, especially on the ocean side, and bar otropic pressure gradient inside the inlet. In addition, inside the inlet, baroclinic pressure gradient turned from positive to negative due to the completely outward residual flow. The system started to return to prestorm conditions when the winds started to relax in the next 24 hours. The system fully recovered in almost 12 days after the first changes within the system, with the exception of baroclinic pressure gradient which was modified due to the increased freshwater input. Momentum Balances during T ropical Depression Ivan The momentum balance did not show much change during Ivan. The balance of the system is still primarily between the barotropic pressure gradient and the nonlinear advection, since the wind was never strong around the region during I van. During the second landfall of Ivan, the cross channel baroclinic pressure gradient changed slightly due to the freshwater input into the system and acted together with the nonlinear advection inside the inlet to balance the barotropic pressure gradient. This momentum balance persisted until Hurricane Jeanne started to affect the region.

PAGE 57

57 Momentum Balances during Hurricane Jeanne Before H Jeanne made landfall, wind stress in both along and cross channel directions increased inside the Tolomato River and inside the inlet ( Figure 3 6 a and Figure 3 7 a), due to winds blowing from northeast with speed of 7 m/s. While the along shore bottom friction did not change significantly, the cross channel bottom friction increased, especially on the ocean side of the inlet around the tips of the headlands due to the southerly along shore currents. Along channel baroclinic pressure gradient continued to persist on the east shore of ICWW without significant change in magnitude. However, a positive cross channel baroclinic pressure gradient formed along the deep navi gation channel inside the inlet due to the change in residual flow structure. Right after the landfall of H Jeanne, the wind changed to southeasterly with speed up to 24 m/s (maximum wind during the simulation period) because of the hurricanes northward m ovement towards the region after the landfall. In the along channel direction, the wind stress was still very significant inside the Tolomato River and parts of the Matanzas River ( Figure 3 6 b). The along channel bottom friction increased mostly along the deep navigation channel inside the inlet. The barotropic pressure gradient balanced both the wind stress and the bottom friction in the along channel direction. In the cross channel direction, wind stress created a negative flux over the entire region ( Figure 3 7 b), while bottom friction increased significantly on the ocean side, balancing the wind stress. The wind stress is balanced by barotropic pressure gradient inside the inlet. A short summary of changes to the residual circulation with respect to the other forces during the simulation period i s given in Table 33

PAGE 58

58 Time V ariati on of M omentum B alances at a S tation I nside the I nlet Snapshots of residual momentum balance over the region surrounding SAI provide excellent information for understanding the causes for changes in residual circulation during storm season. However, since these snapshots are not continuous in time, it is interesting to understand the continuous temporal variation of each term in the m omentum equations. For this purpose time series of each term in the momentum equations and their total at a selected location inside the inlet are given in Figure 3 8 The time series show that the system was in balance with practically no variation in horizontal diffusion and local acceleration throughout the simulation period. In both along and cross channel directions the momentum balance is between the barotropic pressure gradient and nonlinear advection terms, with considerable contribution from the cross channel baroclinic pressure gradient. Wind stress and bottom friction changed slightly during hurricanes. The cross channel baroclinic pressure g radient changed over a longer time period, especially after the hurricanes, compared to the wind stress and bottom friction. River D ischarge and P recipitation River discharge and precipitation are two other external forces that might change the residual f low structure and/or the residual momentum balance. NCB has many small creeks that bring in freshwater, but discharges for most of these streams are very small to affect the entire system. As shown in Figure 3 9 during Hurricanes Charley and Frances, the precipitation increased significantly followed by a substantial increase in the river discharge. Although the system received significant amount of freshwater, it did not cause a significant variation in the residual flow structure. On the other hand, the baroclinic pressure gradients increased significantly, particularly inside the Tolomato

PAGE 59

59 River in the along shore direction and along the deep ICWW insi de the inlet in the crosschannel direction ( Figure 3 10). Similarly, time series for the baroclinic pressure gradient ( Figure 3 8 ) at a station inside the inlet show two peaks coinciding with the increased precipitation and river discharge ( Figure 3 9 ). Vorticity and Evolution of Residual E ddies Residual eddies forming around the SAI area shift or vanish during strong wind events. Their evolution is highly dependent on the wind speed, wind direction and the nonlinear advection. Before Hurricane Frances, during a typical tidal cycle, eddies around the region were stable with strong positive vorticity for the eddy to the south and with strong negative vorticity for the eddy to the north of th e inlet on the estuary side ( Figure 3 12a). 24 hours before H Frances made landfall, the northeasterly winds disrupted the eddies on the ocean side, c reated southerly along shore current and forced inflow through the inlet due to Ekman transport ( Figure 3 12 b). Eddies inside the inlet were intact, but weaker due to the wind stress increasing inside the inlet. Similarly the vorticity strength for both eddies inside the inlet reduced due to a more uniform inflow through the inlet. During the landfall of H Frances, the winds were easterly and the along shore current on the ocean side became northerly due to the Ekman transport created by easterly winds and the relaxing wind induced setup in the along shore direction ( Figure 3 12c). Eddies inside the inlet were mostly disrupted when the wind stress overwhelmed the nonlinear advection and the strength of vorticity became very small. However, as a result of strong nonlinear advection and outflow through t he inlet, another eddy with strong vorticity was formed at the mouth of the inlet. Almost 24 hours after the landfall of H Frances, the winds became southeasterly and the along shore northerly current

PAGE 60

60 became much stronger ( Figure 3 12d). Eddies inside the inlet were completely vanished, and flow through the inlet became complete outflow. The eddy at the mouth of the inlet was intact, but reduced in size with stronger outflow through inlet. The wind speed reduced and the winds turned from southeasterly to southerly in 8 hours causing the along shore currents as well as the flow through the inlet to slow down ( Figu re 3 12e). However, the wind stress both inside and the ocean side of inlet was strong enough to prevent the reformation of the eddies. The eddy at the mouth of the inlet expanded towards the ocean side with the slowing along shore currents and the strength of vorticity reduced due to the slower velocities through the inlet. In the following 48 hours, the wind stress reduced significantly and with the nonlinear advection becoming dominant agai n, the eddies inside the inlet reformed. The eddy at the mouth of the inlet moved towards the ocean side and expanded to form the north eddy on the ocean side. The eddies and vorticity inside the system recovered to their pre storm conditions 12 days after the initial changes in the system occurred ( Figure 3 12 f). Conclusions The impacts of four major hurricanes (Charley, France, Ivan, and Jeanne) in 2004 on the residual circulation and momentum balance inside the subtropical North Coastal Basin and St. Augustine Inlet were investigated using a three dimensional hydrodynamic model, CH3D. The model simulated residual tidal circulation around the inlet show formation of four residual eddies around the barrier islands: two eddies inside the estuary with speeds up to 20 cm/s, and two weaker eddies on the ocean side with speeds up to 10 cm/s. Model results confirm that formation of residual eddies is caused by the nonlinear advection as shown by Tee (1977) and Murty (1980). In addition, the

PAGE 61

61 model results indicate that the cross sectional residual flow observed by Webb et al. (2007) was part of the two threedimensional residual eddies formed inside the inlet. However, since the structure of the residual eddies is highly dynamic, the cross sectional residual flow observed by Webb et al. is both temporally and spatially varying. Although the variations to the spatial structure of the residual circulation during each hurricane showed some differences, the general patterns are similar during all the storms. The approaching hurricanes first disrupted the two eddies on the ocean side and the along shore flow became unidirectional towards south. During or right before a hurricanes landfall, the residual flow became complete inflow throughout the mouth of the inlet due to the changing wind direction. After landfall, changing wind direction forc ed the residual flow at the mouth of the inlet to reverse to a complete outflow, in addition to the completely reversed northerly along shore flow. The system returned to its pre storm conditions after the winds and the barotropic pressure gradient relaxed. Variations in residual flow structure lasted up to 12 days during Frances (a large hurricane), but only 12 days for Charley (a compact and fast hurricane).This difference in the recovery time of the residual flow is also found in the residual salinity distribution. The model results reveal that the momentum balance during a regular tidal cycle is between barotropic pressure gradient and nonlinear advection. During hurricanes, as a result of strong winds, the wind stress around the region gained strength and was a significant part of the momentum balance. Baroclinic pressure gradient was especially important in the cross channel direction with considerable increase after the hurricanes.

PAGE 62

62 a) b) c) Figure 31 Study area, (a) St. Augustine Inlet; (b) Northern Coastal Basin (NCB); (c) Tracks for four major hurricanes in 2004; (d) CH3D modeling domain on the east coast of Florida

PAGE 63

63 a) b) Figure 32 Cross sectional residual circulation obtained in a rotating elongated rectangular basin for moderate friction case ( =0.5) including the rotational forces obtained with (a) analytical model (adopted from Winant 2008) ; (b) CH3D simulation; at four equally spaced cross sections (x=0,0.25,0.5,0.75) starting from the mouth of the channel (x=0) and going towards the closed end (x=1) of the channel. Both models used constant eddy viscosity.

PAGE 64

64 a) b) Figure 33 Spatial pattern of 34 hour low pass filtered vertically averaged residual flow pattern around St. Augustine Inlet region during a regular tidal cycle calculated (a) with the nonlinear advection terms; (b) without the nonlinear advection terms 2 5 c m / s 2 5 c m / s

PAGE 65

65 a) b) Figure 34 Spatial pattern of 34h low pass filtered vertically averaged residual flow patterns around St. Augustine Inlet region. (a) Right before Hurricane Frances landfall (09/04/2004 12:00 UTC); (b) When Hurricane Frances moves across Florida (09/06/2004 12:00 UTC) 2 5 c m / s 2 5 c m / s

PAGE 66

66 a ) b) Figure 35 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region during a regular tidal cycle: (a) The along channel momentum balance; (b) the crosschannel momentum balance. B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 67

67 a) b) Figure 36 Spatial pattern of 34h low pass filtered along channel momentum balance around St. Augustine Inlet region (a) Right before Hurricane Frances landfall (09/04/2004 12:00 UTC); (b) When Hurricane Frances moves across Florida (09/06/2004 12:00 UTC) B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 68

68 a) b) Figure 37 Spatial pattern of 34h low pass filtered crosschannel momentum balance around St. Augustine Inlet region. (a) Right before Hurricane Frances landfall (09/04/2004 12:00 UTC); (b) When Hurricane Frances moves across Florida (09/06/2004 12:00 UTC) B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 69

69 Figure 38 Non dimen sional time series of 34h low pass filtered along channel (solid line) and cross channel (dashed line) momentum flux terms at a selected station inside the inlet for approximately two month period between 08/06/2004 and 09/31/2004 covering four major hurr icanes in 2004.

PAGE 70

70 Figure 39 River discharge (Pellicer Creek) and precipitation (Domenico Circle) around St. Augustine Inlet during the simulation period

PAGE 71

71 a) b) Figure 310. Momentum balance occurring after hurricanes with considerable precipitation (a) along channel; (b) cross channel B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 72

72 a) b) c) d) Figure 311. Variation of cross sectional residual flow at the St. Augustine Inlet at three different transects (a) location of transects; (b) transect 1; (c) transect 2; (d) transect 3. Distance for each transect is measured from the south shore of the inlet.

PAGE 73

73 a) b) c) d) e) f) Figure 312. Evolution of residual eddies, streamlines and vorticity at different phases of Hurricane Frances (a) during a typical tidal cycle before Hurricane Frances; (b) 24 hours before the landfall; (c) during landfall; (d) 24 after landfall; (e) 32 hours after l andfall; (f) 12 days after first variations to the residual flow structure V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s V o r t i c i t y ( 1 / s )1 5 1 0 5 0 0 5 1 1 5 2 5 c m / s 1 0 m / s

PAGE 74

74 Table 3 1 Features of four hurricanes affecting the region in 2004. All of the hurricane parameters given in the table represent the conditions when the storm was at its minimum distance to the St. Augustine Inlet. Intensity Max. Wind Speed (m/s) Max. Wind Radius (km) Min. Distance (km) Charley Cat 1 9 5 40 Frances Cat 2 23 30 325 Ivan Tropical Depression 6 60 280 Jeanne Cat 3 24 25 325

PAGE 75

75 Table 32 Summary of changes to residual flow structure and their cause Changes to the residual flow Caused by* Charley and Ivan 1. Ocean side eddies disrupted 1. Wind from northeast Frances and Jeanne 1. All eddies disrupted 2. Complete inflow 3. Complete outflow 1. Northeast, southeast winds 2. Ekman transport due to northeast and east winds 3. Ekman transport due to southeast and s outh winds Freshwater input into the system during hurricane was not effective to change the residual flow structure

PAGE 76

76 Table 33 Summary of changes to residual momentum balance and their cause Changes to the momentum balance Caused by Charley, Ivan, Frances and Jeanne 1. Increased wind stress and barotropic p.g. 2. Increased bottom friction* 1. Increased wind speed 2. Increased residual flow velocity After hurricanes 1. Increased baroclinic pressure gradient* 1. Increased freshwater input during hurricanes The balance was still between barotropic pressure gradient and nonlinear advection.

PAGE 77

77 CHAPTER 4 EFFECT OF TROPICAL C YCLONES ON RESIDUAL CIRCULATION AND MOMENTUM BALANCE IN A SUBTROPICAL ESTUAR Y AND INLET OBSERVATION AND SIMU LATION3 Abstract A three dimensional hydrodynamic model, CH3D and observations are used to investigate the effects of tropical storms Fay and Hanna (2008) on the spatial structure of residual circulation and momentum balance inside a subtropical estuarine system Northern Coastal Basin (NCB) and the St. Augustine Inlet (SAI). Simulated 3D velocities are successfully verified with the observed velocities inside the SAI. During a typical tidal cycle, residual circulation showed formation of four residual eddies; two strong eddies inside and two weaker eddies outside of the inlet. During Tropical Storms Fay and Hanna, along shore co mponent of the northeasterly winds disrupted the ocean side eddies first and forced along shore southeasterly current. Due to relatively weak local wind effects, residual flow at the mouth of the inlet never became complete inflow. However, following TS Fays landfall, flow inside the inlet became complete outflow due to relaxing wind and remote wind effect created by the along shore component of the southeasterly wind. The residual circulation returned to its prestorm condition in a short time (4 days) due to storm duration. Model results revealed that the cross sectional residual flow pattern observed at the inlet mouth is highly variable with time and location of the cross section. Comparison of terms in the horizontal momentum equations showed that the balance is between the barotropic pressure gradient and nonlinear advection. Although the wind stress and bottom friction briefly become significant in the 3 To be submitted as Tutak, B., Y.P. Sheng and V. Paramygin: Effect of Tropical Cyclones on Residual Circulation and Momentum Balance in a Subtropical Estuary and Inlet Observation and Simulation

PAGE 78

78 balance, barotropic pressure gradient and advection were still and order of magnitude larger. Barocl inic pressure gradient was weak despite the increase in precipitation and river discharge. Introduction Residual circulation is defined in different ways as either the average of tidal flow over a tidal cycle (Fischer et al. 1979) or net movement of water over a tidal cycle (Cheng and Casulli 1982). Residual circulation can be created due to physical characteristics of the estuary and inlet including bathymetry (Kjerfve 1978), or geometry of the channel (Li et al. 2008). In addition, the residual circulatio n could be created as a result of different forces like bottom friction (Winant 2008), Coriolis (ValleLevinson et al. 2003), density gradient (Fischer 1976), nonlinear advection terms (Murty et al. 1980; Pingree and Maddock 1977) and wind forcing (Janzen and Wong 2002). It is important to have a sound understanding of the residual circulation because it affects the long term dispersal of pollutants, nutrients and larvae, estuarine flushing times or sediment transport (Prandle 1978; Sheng et al. 2008). Stud ies on the residual oceanestuary exchange flow at the mouth of a channel or an inlet have been conducted by many researchers. Li and ODonnell (1997) studied the effects of laterally varying bathymetry (v shaped) for shallow estuaries and found that the r esidual exchange pattern was inflow over the shoals and outflow in the deep channel at the mouth of the rectangular channel. In a similar study, Li and ODonnell (2005) used a two dimensional analytical model to study the residual circulation and channel l ength relation. Li and ODonnell identified the residual exchange pattern in short channels, which are characterized by standing wave, as inflow through the deep

PAGE 79

79 channel and outflow over the shoals, opposite to that in long channels characterized by progre ssive waves. Webb et al. (2007) through a field study at St. Augustine Inlet, found supporting evidence of the residual flow theory developed by Li and ODonnell (2005). The crosssectional residual flow structure measured by Webb et al. showed inflow thr ough the deep channel and outflow over the shoals ins ide the St. Augustine Inlet. Moreover, Webb et al. showed through order of magnitude analysis that the force balance inside the inlet is dominated by nonlinear advection and barotropic pressure gradient. Sheng et al. (2008) simulated the flow measured by Webb et al. (2007). Along estuary (local) and alongshore (remote) wind forcing might create variation in the residual estuarine circulation Variation in residual circulation might be cau sed by the local wind stress acting directly over the estuary, or remote wind forcing might create a change in subtidal pressure gradients at the entrance of the estuary (Wong, 1994). Subtidal currents tend to show a more unidirectional flow pattern throughout the water c olumn when remote winds are dominant ( Garvine 1985), while a vertically sheared flow structure is observed with local wind effect (Wang 1979). A significant variation in the residual flow at the mouth of the Chesapeake Bay in response to Hurricane Floyd was shown by ValleLevinson et al. (2002) utilizing observed data at several stations at the mouth of the bay. The flow structure that was inflow over the shoals and outflow in the deep channel prior to the Hurricane Floyd became a complete outflow throug hout the bay entrance due to the effect of wind and freshwater discharge. Smith (2001) studied seasonal scale variations in transport patterns due to the changes in sea level change and freshwater discharge. Smith found

PAGE 80

80 that a flood dominated tidal inlet m ight become ebb dominated due to t he secondary forcing functions. Although analytical models can provide significant insight for understanding the residual circulation in idealized estuaries and inlets with simple geometry and bathymetry, they cannot be readily applied to real estuaries and inlets with complex bathymetry and geometry such as the NCB and SAI. Therefore, studies in residual circulation in real estuaries and inlets have been conducted by using two or three dimensional numerical hydrodynamic models. Tee (1977) used a two dimensional vertically integrated numerical model to understand the behavior of the residual circulation inside the Minas Basin in Bay of Fundy and found residual eddy formation was a result of the nonlinear advective terms. Hench and Luettich (2003) studied the near inlet momentum balance inside Beaufort Inlet (North Carolina) by a twodimensional vertically integrated numerical model, and found that the balance during flood phase was between barotropic pressure gradient and adv ection, and between advection and bottom friction during the ebb phase. Li (2006) examined the residual eddies generated in tidal channels using both an analytical and a three dimensional hydrodynamic model. Li found that for complex bathymetries residual eddies might become more important than the inout type of residual exchange flows and the structure of these residual eddies does not change with tidal range fluctuations (e.g. spring and neap). Table 41 is a summary of the above mentioned studies: This Study The main objective of this study is to understand the residual circulation and momentum balance inside the St. Augustine Inlet (SAI) ( Figu re 4 1 b) and the Northern Coastal Basin (NCB) estuarine system ( Figure 41 a) during a typical tidal cycle as well

PAGE 81

81 as to understand changes to the residual circulation and momentum balance during two tropical storms in 2008 (Fay and Hanna) by using a threedimensional hydrodynamic model CH3D and observed velocities inside the SAI. The simulation period covers two tropical storms Fay in August and Hanna in September 2008 ( Figure 41 c). The model domain includes the entire NCB to enable a comprehensive understanding of the system wide hydrodynamic processes. NCB is an estuarine system that extends from Pablo Creek (east of Jacksonville, Florida) at the north to the Ponce de Leon Inlet (south of Daytona Beach, Florida) at the South, and connected all the way from n ort h to south through the Intracoastal Waterway (ICW W ) ( Figure 41 a) A comprehensive understanding of the hydrodynamics inside the NCB during extreme events is important because the area is susceptible to drastic changes due to extreme events. For example, in 2004, four major hurricanes that made landfall in Florida affect ed this region significantly (Tutak and Sheng 2010a; 2010b). Salinity in some regions of NCB reduced to zero and the system returned to its prestorm conditions almost after two months from the first hurricane (Hurricane Charley). Similarly, in 2008 two tr opical storms affected the region, although one of them never made landfall in Florida. Therefore, understanding the spatial and temporal variation of the residual circulation and momentum balance around the SAI and adjacent areas during regular tidal cycl e and episodic storm events will help understanding the transport mechanisms within the system. This study aims to answer the following questions and provide a more comprehensive understanding of the phenomenon observed in previous studies: What is the spa tial distribution of residual circulation around the region? Which forces dictate the residual momentum balance around the region? How do tropical storms affect the patterns of residual circulation and the momentum balance around the region?

PAGE 82

82 Study Area and Methods The NCB system is connected to the Atlantic Ocean through 3 inlets : St. Augustine, Matanzas and Ponce de Leon. SAI is the northernmost inlet within the system and it is the widest and deepest inlet of the system compared to the other two inlets. T idal range changes approximately from 1. 2 m at the south to 1. 5 m at the north All three inlets within the system are tidally dominated inlets and t he flow within the NCB system is dominated by the tides mostly through the SAI The volumetric flow rate thro ugh SAI is approximately 5 times larger compared to Matanzas and Ponce de Leon Inlets. The estuary, including the ICW W is very shallow (with a minimum of 0.3 m and an average of 2.7 m) and narrow. The system consists of 13 small rivers, creeks and tributar ies that bring freshwater in to the estuarine waters For hydrodynamic simulations, the C urvilinear Grid H ydrodynamics in 3D (CH3D) model (Sheng 1987, 1990; Sheng et al. 2010a) is used. CH3D is a robust hydrodynamic model that can simulate circulation driv en by tide, wind, and density gradients. T he model is also capable of importing large scale and hurricanescale wind fields and handling flooding and drying of shorelines for storm surge simulations. In addition, CH3D is chosen for this study because of it s ability to (1) accurately resolve t he complex shoreline and geometry of NCB with boundary fitted curvilinear grid; (2) accurately resolve the rapid changes in bathymetry using the terrain following vertical grid; (3) accurately represent the v ertical turbulent mixing with a robust turbulence closure model (Sheng and Villaret 1989). The model has been extensively used and verified for similar Florida estuaries including GTMNERR (Sheng et al. 2008), Indian River Lagoon (Sheng and Kim 2009), Pensa cola Escambia Bay (Sheng et al. 2010b),

PAGE 83

83 and Charlotte Harbor (Kim et al. 2010). Details of the equations and boundary conditions can be found in Sheng et al. (2008) and Sheng et al. (2010a). The curvilinear boundary fitted grid used for this study is a ver y high resolution grid with a minimum grid spacing of 15 m inside the estuary for resolving the narrowest navigation channels. The grid ( Figure 41 d) is composed of a total of 37250 water and 22640 floodable cells. Eight vertical sigma layers are used to handle the bathymetry changes from very shallow inside the estuary to very deep offshore and to capture the threedimensional velocity structure of the residual circulation. Time step used for the numerical simulations is 10s. Moreover, simulations make use of the flooding anddrying feature to accurately account for the inundation due to tides or storm surge caused by tropical storms. Boundary and initial conditions req uired for s imulating the hydrodynamics during extreme events are derived from observed data at stations scattered throughout the system and/or from regional scale numerical model outputs T idal boundary condition for CH3D is obtained from ADCIRC tid al data base for the East Coast (Mukai et al. 2001) as t idal constituents at the offshore boundary of CH3D. Wind and atmospheric pressure during nonstorm time periods are obtained from the Navy Operational Global Atmospheric Prediction System (NOGAPS) model (Barr on et al. 2007) output. During storm periods, 3 hourly or 6hourly snapshots of the storm winds are obtained from the NOAA (National Oceanic and Atmospheric Administration) Hurricane Research Division (HRD) H*Wind 10m rea nalysis winds (Powell et al. 1998), and blended with the background NOGAPS winds (Paramygin, 2009) Since H*Wind snapshots only contain wind information, the a tmospheric pressure is obtained using the Holland B (Holland

PAGE 84

84 1980) analytical wind model. To simulate baroclinic circulation, CH3D uses river discharge and precipitation data. Discharge data available from USGS, is included into the model as boundary condition at the upstream of river s. Precipitation data is obtained from St. Johns River Water Management District (SJRWMD). In additi on, a bottom mounted Acoustic Doppler Current Profiler (ADCP) was deployed ( Figure 41 b) inside the deep navigation channel (at approximate 15 m) to measure the verti cal velocity profile over the entire water column at every 0.5 m at a ping rate of 1.5 s, averaging 400 samples per ensemble. The deployment was from July 18 to September 19, 2008 and velocity observations are used to verify the model results. Model Verifi cation with ADCP observations Before simulating the residual circulation and momentum balance during TS Fay and TS Hanna, simulated velocities are compared with observed velocities. The simulated vs. measured vertically averaged velocities, as shown in Figure 4 2 agree very well. The average normalized RMS error for simulated vertically averaged velocities was less than 8%, while it was approximately 10% for the simulated threedimensional velocities in both directions of the flow. More detailed model verification for the 2008 tropical storms as well as the 2004 season is given in Tutak and Sheng (2010a). Spatial Patterns of Residual Circulation During One Typical Tidal Cycle The spatial structure of the residual flow around the SAI region shows multiple eddies forming inside and outside of the inlet during typical tidal conditions ( Figure 4 3 a). Two eddies formed inside the inlet were stronger (speed ~1520 cm/s) and more

PAGE 85

85 obvious than the ones formed outside of the inlet (speed ~510 cm/s). The two eddies inside the inlet are located towards the north and south of the inlet. The north eddy rotates clockwise enhancing the flow into the Tolomato River, while the south eddy rotates counterclockwise enhancing the flow towards Matanzas River and Salt Run. On the ocean side, the eddies rotate in the opposite direction compared to the eddies inside the inlet. The north eddy rotates counterclockwise, while the south eddy rotates clockwise. The model results confirm the cross sectional residual flow observed along a transect at the mouth of the inlet by Webb et al. (2007). However, model results also revealed that the cross sectional residual flow pattern was a part of the threedimensional residual eddies ( Figure 4 3 cd) observed inside the inlet. In fac t, the model results show that the cross sectional residual flow varies significantly with the location of the transect ( Figure 4 12). Fischer et al. (1979) speculated that the formations of eddies might depend on the shape and geometry of the channel and the flow separation that occurs around the headlands. Zimmerman (1978), using an analytical model, showed eddy formation was due to the interaction of tidal flow with irregular bathymetry. With numerical experiments Tee (1977) showed that a significant portion of the residual eddy formation was caused by the nonlinear advection. Similarly, our model results for the residual circulation with and without the nonlinear advection for the SAI region showed that the generation of the residual circulation and eddies around the region was caused by the nonlinear advection ( Figure 4 3 b). Changes to the Residual Flow Pattern during Tropical Storm Fay During the progression of Tropical Storm (TS) Fay through Florida, the residual flow pattern showed several changes for limited time periods depending on the direc tion and the speed of the storm winds and how the estuary responded to the overall forces.

PAGE 86

86 The residual flow pattern started showing signs of changes ( Figure 4 4 a) when TS Fay emerged to the Atlantic Ocean around Melbourne Beach approximately 200 km south of the region. The wind was blowing persistently from northeast with approximately 10 m/s reaching up to 18 m/s during the landfall. During this time t he eddies on the ocean side of the inlet completely vanished ( Figure 4 4 a) due to the along shore (northerly) component of the wind stress. The residual flow on the ocean side was along shore towards south with speed up to 20 cm/s. Part of the flow still entered the inlet and moved into the estuary instead of creating an eddy. Analysis showed that along shore component of the wind stress has a high corr elation (R= 0.75) with the rate of change of water level ( / t) inside the inlet, which is considered the proxy for the remote wind forcing (Janzen and Wong 2002). This remote wind effect forced a unidirectional Ekman transport into the inlet. The two eddies inside the inlet preserved their shapes and locations and, with the extra flow coming from the ocean side, both eddies strengthened. This condition persisted for approximately 6 hours right until the TS Fay made landfall around Flagler Beach. After the landfall, the flow inside inlet went through a dramatic change ( Figure 4 4 b) in the following 36 hours. The eddies inside the inlet lost most of t heir strength but still existed. However, the flow through the mouth of the inlet became completely out of the estuary. When the wind direction turned from northeasterly to southeasterly, the along shore component of the wind stress forced a unidirectional Ekman transport out of the inlet acting as a proxy for remote wind forcing. The south eddy gained strength by flow coming through the Matanzas River and reaching into the inlet. In addition to its remote effect, the along shore component of the wind acted as a highly correlated (R =

PAGE 87

87 0.96) proxy to local wind forcing due to the direction of ICWW being parallel to the shoreline. The strong outflow from the inlet helped the eddies outside the inlet to reform, but these eddies were still very weak. The residual circulation pattern inside the inlet returned to its prestorm conditions approximately 4 days after the residual flow pattern started to change. Effect of cross shore component of the wind stress in changing the residual circulation was very small, whic h was confirmed by its poor correlations to all related parameters. The variations in the residual flow structure during TS Fay show similarities to the variations obtained by Tutak and Sheng (2010b) during the four major hurricanes in 2004. However, the r ecovery time of residual flow to its prestorm condition was shorter (4 days) during TS Fay, especially against the recovery time (12 days) during H Frances in 2004. In addition, due to the strength and directionality of the wind during TS Fay, a complete inflow condition was never reached. Changes to the Residual Flow Pattern during Tropical Storm Hanna Although TS Hanna traveled over the Atlantic Ocean without making landfall, it still affected the region and modified the residual flow pattern around the St. Augustine Inlet. The storm at its closest point was 200km away from the Florida coast. The winds from the storm reached 12 m/s around the inlet and caused the eddies on the ocean side to disrupt and become a southerly along shore current ( Figure 4 5 ) as it happened during TS Fay, followed by an increase in the strength of the residual eddies inside the inlet. Strengthening of the two inside eddies mainly happened due to the extra flow that came from the ocean side and enhanced the inflow. The local and remote wind forcing was also similar to that observed during TS Fay. The system recovered to its prestorm condition in approximately 24 hours.

PAGE 88

88 Spatial Patterns of Momentum Balance During One Typical Tidal Cycle During a typical tidal cycle, the spatial structure of the momentum balance inside the inlet is determined by two forcing functions. The most dominant forcing function inside the inlet is the bar otropic pressure gradient and it is balanced by the nonlinear advection terms ( Figure 4 6 a b) in both along channel (approximately 15 counterclockwise from E W direction) (Equation 1) and cross channel (Equation 2) directions. Along with the nonlinear terms, bottom friction is significant especially at the mouth of the inlet and over tidal flats inside the inlet (ebb shoals), where the bathymetry is fairly shallow. Coriolis force is almost nonexistent, except inside the inlet on the north shore in the along channel momentum balance, where the shoreline shows a significant curvature. Although Coriolis and bottom friction seem to be important at specific locations, most of the time, their magnitude is at least an order of magnitude smaller than the barotropic pressure gradient and the nonlinear advection terms. Both the horizontal diffusion and local acceleration are negligible throughout the simulation period and hence are not plotted. + ( / ) + ( / ) + + + + = 0 Eq. ( 4 1) + ( / ) + ( / ) + + + + + = 0 Eq. ( 4 2) Changes in Momentum Balance during Tropical Storm Fay With the arrival of TS Fay most noticeable change in momentum balance happened with the significant wind stress caused by strong storm winds ( Figure 4 7 a -

PAGE 89

89 Figure 4 8 a). The alongchannel wind stress increased inside the Tolomato River, mouth of Matanzas River, and inside the inlet. The cross channel wind stress increased significantly inside the entire inlet. Wind at the moment was blowing from northeast with approximately 10 m/s reaching up to 18m/s during landfall. Inside the Tolomato River, along with the wind stress, bottom friction showed a small increase in momentum flux in the along channel direction. Bottom friction in cross shore direction responded to the increase in wind stress, creating a negative momentum flux, especially on the ocean side around the shoreline. The momentum flux created by the wind stress and bottom friction was balanced mostly by the barotropic pressure gradient. Right after land fall the wind stress started relaxing. In along channel direction it was still significant inside the Tolomato River, and in the cross channel direction, momentum flux changed direction ( Figure 4 7 b Figure 4 8 b) with the movement of the storm fr om the east of inlet to the west of inlet, which changed the wind direction from northeast to southeast. The wind stress was balanced by the barotropic pressure gradient with bottom friction losing strength. After the landfall, with increased precipitation and river discharge around the region, the baroclinic pressure gradient started showing small increase inside the Tolomato River and at the mouth of the inlet in both along and cross shore directions. The system recovered to its prestorm conditions after almost 4 days. The wind stress completely relaxed along with the balancing barotropic pressure gradient and bottom friction inside the Tolomato River. The only exception to the recovery happened with the baroclinic pressure gradient inside the Tolomato Ri ver and inside the inlet. The baroclinic pressure gradient existed mainly due to the heavy precipitation and river discharge during and after the TS Fay ( Figure 4 11).

PAGE 90

90 Changes in Momentum Balance during Tropical Storm Hanna During TS Hanna, the along channel momentum balance throughout the region did not show significant variability compared to that during a typical tidal cycle. For most of the region, t he balance is still between the barotropic pressure gradient and the nonlinear advection terms. For the cross channel momentum balance, the effect of wind stress and bottom friction created a considerable momentum flux. When the storm was at its closest po int to the Florida coast (200 km offshore), wind stress and baroclinic pressure gradient showed very little increase in the along channel direction ( Figure 4 9 a). In the cross shore direction, the response of the system was similar to the one seen during the TS Fay ( Figure 4 9 b). The momentum flux due to wind stress increased inside and outside of the inlet, while bottom friction increased to balance the wind stress. The recovery of the system after TS Hanna happened very quickly. In less than 24 hours, the system returned to its prestorm conditions. Time V ariation of M omentum B alance at a Station Inside the Inlet Snapshots of momentum balance at different times show the variation of the momentum balance over the given area. However, snapshots alone do not show the order of magnitude differences between the different forcing terms. For this purpose, we selected a station inside the inlet and plotted the time series of the complete momentum balance in along channel and cross channel directions ( Figure 4 10 ). The location of the station was selected at the location of ADCP observations given in Waterhouse et al. (2010). The momentum flux obtained from the model results was completely at balance at the selected location during the entire simulation period. As in the snapshots, the main balance in both the along channel and cross channel directions was between the barotropic pressure gradient and the nonlinear advection terms. In addition, these tw o

PAGE 91

91 terms were almost an order of magnitude greater than all the other terms. During TS Fay (~day 235), the wind stress and bottom friction at this location showed some increase, but they were still very small compared to the two main forces. Right after TS Fay, baroclinic pressure gradient in the cross channel direction (directed along ICWW), showed a substantial increase in magnitude and contributed to the balance of two significant forces until the end of TS Hanna. Since the cross channel direction indicat es the changes in the ICWW, the sudden dramatic change in baroclinic pressure gradient can be attributed to the increase in precipitation and river discharge right after TS Fay ( Figure 4 11). Coriolis force, which was almost nonexistent in the along channel direction, generated a cross channel momentum flux that was comparable to the bottom friction during the TS Fay. Horizontal diffusion and local acceleration in both along channel and cross channel directions were nonexistent. In addition, Waterhouse et al. (2010) calculated the momentum balance for the same location using observed data around the region in a similar time period and also suggested that the balance was between barotropic pressure gradient and nonlinear advection. Conclusions Th e spatial distribution of residual circulation and momentum balance in the subtropical Northern Coastal Basin (NCB) and St. Augustine Inlet (SAI) during two tropi cal storms in 2008 is calculated using a threedimensional numerical model. The simulated vertically integrated velocities agree very well with those obtained by a bottom mounted ADCP inside the SAI. During a typical tidal cycle, the spatial structure of t he residual circulation show that four eddies form around the inlet, two strong eddies (speed ~1520 cm/s) inside the inlet and two relatively weaker eddies (speed ~510 cm/s) form outside of the inlet. The two counter rotating eddies inside the inlet create a

PAGE 92

92 crosssectional residual flow pattern of outflow at the shallow north section and inflow at the deeper south section at the mouth of inlet consistent with the cross sectional observation by Webb et al. (2007). However, since the residual eddies inside the inlet show highly spatially variable rotating motion, the cross sectional residual flow along different transects can be very different. The analysis of the momentum balance inside the inlet and surrounding area showed that the balance was predominant ly between the barotropic pressure gradient and nonlinear advection terms. In addition, model results with and without the nonlinear terms showed that the residual circulation was largely due to the nonlinear advection, as discovered by others. The residual eddies generated on the ocean side during typical tidal cycles were disrupted by the along shore component of the strong northeasterly winds before landfall and forced to a southerly along shore current. The residual flow structure after TS Fa y became complete outflow for almost a day with the relaxing winds and the Ekman transport created by the along shore component of the southeasterly wind, which acted as a proxy for remote wind forcing. The total amount of time during which the residual ci rculation was from the typical tidal residual circulation was approximately 4 days starting from almost a day before TS Fay made its landfall. With the effect of stronger wind, the wind stress gained strength, and balanced by the combination of barotropic pressure gradient and bottom friction. In addition, with the effect of river discharge and considerable precipitation, the baroclinic pressure gradient showed slight increase. Although some of the terms in the momentum balance strengthened, overall balance was not significantly modified from the balance between barotropic pressure gradient and nonlinear advection terms. Both the barotropic pressure gradient and the

PAGE 93

93 nonlinear advection were generally an order of magnitude higher compared to the other terms t hroughout the simulation period.

PAGE 94

94 a) b) c) Figure 41 Study area, (a) Northern Coastal Basin (NCB); (b) St. Augustine Inlet; (c) Tracks for TS Fay and TS Hanna in 2008; (d) CH3D modeling domain on the east coast of Florida

PAGE 95

95 a) b) Figure 42 Comparison of observed (.) and simulated (solid) vertically averaged velocity time series during 2008 simulation time period in a) east west; b) northsouth directions J u l i a n D a yV e l o c i t y ( c m / s ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 2 0 0 1 0 0 0 1 0 0 2 0 0 O b s e r v e d S i m u l a t e d J u l i a n D a yV e l o c i t y ( c m / s ) 2 3 0 2 3 5 2 4 0 2 4 5 2 5 0 2 5 5 6 0 3 0 0 3 0 6 0 O b s e r v e d S i m u l a t e d

PAGE 96

96 a) b) c) d) Figure 43 Spatial pattern of 34h low pass filtered residual circulation pattern around St. Augustine Inlet region during a regular tidal cycle (a) vertically averaged with the nonlinear advection terms; (b) vertically averaged without the nonlinear advection terms; (c) 3D bottom velocity; (d) 3 D surface velocity. 2 5 c m / s 2 5 c m / s 2 5 c m / s 2 5 c m / s

PAGE 97

97 a) b) Figure 44 Spatial pattern of 34h low pass filtered vertically averaged residual flow patterns around St. Augustine Inlet region (a) During landfall of TS Fay (08/21/2008 19:00 UTC); (b) After TS Fay landfall (08/22/2008 00:00 UTC) 2 5 c m / s 2 5 c m / s

PAGE 98

98 Figure 45 Spatial pattern of 34h low pass filtered vertically averaged residual flow pattern around St. Augustine Inlet region when TS Hanna closest to the Florida Coast (09/05/2008 12:00 UTC) 2 5 c m / s

PAGE 99

99 a) b) Figure 46 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region during a regular tidal cycle in (a) The along channel momentum balance; (b) the cross channel momentum balance. B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 100

100 a) b) Figure 47 Spatial pattern of 34h low pass filtered along channel momentum balance around St. Augustine Inlet region (a) During landfall of TS Fay (08/21/2008 19:00 UTC); (b) After TS Fay landfall (08/22/2008 00:00 UTC) B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 101

101 a) b) Figure 48 Spatial pattern of 34h low pass filtered crosschannel momentum balance around St. Augustine Inlet region (a) During landfall of TS Fay (08/21/2008 19:00 UTC); ( b ) After TS Fay landfall (08/22/2008 00:00 UTC) B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 102

102 a) b) Figure 49 Spatial pattern of 34h low pass filtered momentum balance around St. Augustine Inlet region when TS Hanna closest to the Florida Coast (09/05/2008 12:00 UTC) in a ) along channel; (b) crosschannel directions B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G B o t t o m F r i c t i o n C o r i o l i s W i n d S t r e s s N o n l i n e a r A d v e c t i o n B a r o c l i n i c P G M o m e n t u m F l u x5 0 0 4 5 0 4 0 0 3 5 0 3 0 0 2 5 0 2 0 0 1 5 0 1 0 0 5 0 0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 4 5 0 5 0 0 B a r o t r o p i c P G

PAGE 103

103 Figure 410. Non dimensional time series of 34h low pass filtered along channel (solid line) and cross channel (dashed line) momentum flux terms at a selected point inside the inlet for a month period between 08/15/2008 and 09/15/2008 covering TS Fay (day 235) and TS Hanna (day 250).

PAGE 104

104 Figure 411. Observed river discharge at Pellicer Creek (south of St. Augustine Inlet) and cumulative daily rainfall at Domenico Circle station (12 km south of St. Augustine Inlet)

PAGE 105

105 a) b) c) d) Figure 412. Variation of cross sectional residual flow at the St. Augustine Inlet at three different transects (a) location of transects; (b) transect 1; (c) transect 2; (d) transect 3. Distance for each transect is measured from the south shore of the inlet.

PAGE 106

106 Tabl e 41 Summary of the existing literature Study Method Domain Nonlinear Advection Bottom Friction Tee (1977) 2 D Numerical Model Minas Basin Yes Quadratic with constant coefficient Hench (2003) 2 D Numerical Model Idealized + Beaufort Inlet Yes Manning Chezy Liu (1992) 1 D Numerical model Indian River Lagoon No Quadratic with constant coefficient Li (1997) 2 D Analytical Model Rectangular channel Yes (perturbation) Quadratic by Fourier decomposition Li (2005) 2 D Analytical Model Rectangular channel Yes (perturbation) Quadratic by Fourier decomposition Li (200 6 ) 2 D Analytical Model + Observation Rectangular channel + Wassaw Sound Yes (perturbation) Quadratic by Fourier decomposition Webb (2007) Observed Data St. Augustine Inlet Scaling Analysis Scaling Analysis Valle Levinson (2002) Observed Data Chesapeake Bay Mouth Valle Levinson (2003) 2 D Analytical Model Rectangular Channel No Quadratic with constant coefficient Winant (2008) 3 D Analytical Model Rectangular Channel Yes (perturbation analysis) Quadratic with constant coefficient This Study 3 D Numerical Model + Observation NCB Estuarine System Yes Quadratic with variable coefficient

PAGE 107

107 CHAPTER 5 CONCLUSION In this study, different aspects of the effects of tropical cyclones on the estuarine hydrodynamics were examined at each of the chapters. During tropical cyclones in 2004 and 2008, the salinity inside the system was subjected to significant variations. The changes were generally larger and more prominent close to the mouth of the rivers and creeks due to the increased freshwater discharge associated with the tropical cyclones. Two of these local areas inside the Northern Coastal Basin (NCB) were identified as the Pellicer Creek and Halifax River regions. T he variation of salinity was investigated using surface area of predetermined salinity ranges as: a) low salinity (010 ppt), b) moderate salinity (1020 ppt), and c) high salinity (20 ppt and above). The variation of areas with different salinity ranges was very well correlated with the freshwater input either river discharge or precipitation. In addition, the residual flow rate through each of the three inlets (St. Augustine, Matanzas and Ponce de Leon) was highly correlated with the sharp changes in the area of high or low salinity conditions. For example, in 2004 right before Hurricane Frances made landfall, the winds associated with the hurricane were northeasterly/easterly and were pushing the high salinity ocean water into the system through all thre e inlets. Therefore the area of the high salinity significantly increased. However, when the winds changed direction after the landfall and the wind induced surface setup relaxed along with the increased precipitation and river discharge, the residual flow at the inlets reversed and became complete outflow. This resulted in sharp reduction in the area of high salinity and increase in the low salinity. This shows that strong winds associated with tropical cyclones might enhance or hinder the transport depending on the modified residual

PAGE 108

108 circulation patterns. In addition, variation of surface and bottom salinities as well as the stratification throughout the entire ICWW was analyzed. The analysis of bottom and surface salinities inside the ICWW showed that the flushing of the system was quite slow with up to 15 days around Halifax River at the south and near Pine Island at the northern section of the system as previously shown by Sheng et al. (2008). The transport of salt was towards south inside the Halifax Riv er and around the inlets. High salinity ocean water was moving towards south after entering the system from both St. Augustine and Matanzas inlets due to a southward residual flow created by the higher mean surface elevation at the north of the system. Sim ilarly, the freshwater that was discharged into the system from the rivers and creeks was moving towards south along the ICWW due to mean surface elevation difference between north and south. Although the system went through many cycles of low and high sal inity variations during both 2004 and 2008 simulation periods, the stratification was almost negligible. Only regions with considerable stratification were localized around the mouths of rivers and creeks. The average difference between bottom and surface salinity was generally less than 1 ppt. In addition, during strong winds due to hurricanes or tropical storms, stratification even at the mouth of rivers and creeks almost disappeared due to large vertical mixing by the winds. In addition to the instantaneous hydrodynamic variations during tropical cyclones in 2004 and 2008, the residual hydrodynamic parameters were calculated inside the NCB and around the St. Augustine Inlet. In order to understand the variation in the residual flow during tropical cyclones, first the tidal residual flow was calculated. Tidal residual flow around the St. Augustine Inlet show formation of four residual eddies. Two

PAGE 109

109 of the four eddies form inside the estuary and they were strong with flow speed up to 1520 cm/s. The other two form on the ocean side of the inlet and they were relatively weaker with flow magnitude up to 510 cm/s. Numerical simulations with and without the use of nonlinear advection terms show that the residual circulation around the SAI is almost negligible (~12cm/s) without the nonlinear advection terms. The structure of the residual eddies showed that the cross sectional residual flow along the inlet is highly variable depending on the location of the cross section The cross sectional residual flow along a tr ansect on the estuary side of the inlet shows inflow through the deep channel and outflow over the shallow sections. However, along a transect on the ocean side of the inlet, the cross sectional flow shows inflow through the shoals and outflow through the deep channel. Although each tropical storm or hurricane acted differently on the system, the variation of the residual circulation around the St. Augustine Inlet was quite similar during each of the storms. The change in the system started with the northeasterly winds before the landfall of the storm, which disrupted the two residual eddies on the ocean side and created a uniform along shore southerly flow. In the mean time, inflow through the SAI was forced due to the Ekman transport created by the northeasterly winds and the southerly along shore currents. The two residual eddies inside the estuary persisted. When the winds started to intensify, the residual eddies inside the estuary started to slow down. After hurricanes or tropical storms made landfall, the winds changed direction to easterly followed by southeasterly. As a result of direction change and increase in wind speed, the flow on the ocean side completely reversed and became uniform northerly flow in the along shore direction. Similarly, the flo w inside the estuary as well as the mouth of the inlet reversed and became complete

PAGE 110

110 outflow due to the Ekman transport, and relaxing of the wind induced surface setup inside the estuary. The flow inside the inlet was quite strong with velocities up to 20 cm/s. The recovery time of the system to its prestorm conditions was highly dependent on the strength and the duration of the hurricanes. The recovery of the system was around 12 days during Hurricane Charley (a fast moving compact hurricane), 4 days for TS Fay (a weak, but slow moving storm) and approximately 12 days for Hurricane Frances (a distant, but strong Category 2 hurricane). Similar to the residual flow, residual momentum balance was calculated around the St. Augustine Inlet. During a regular ti dal cycle the residual momentum balance is dominantly between the barotropic pressure gradient and the nonlinear advection terms. Bottom friction is relatively important at several shallow sections of the inlet. During the tropical cyclones the change in t he residual momentum balance was less complicated compared to the residual flow, as the main force contributing or changing the balance was wind stress. When the wind speed increased during storms, the wind stress became important in the residual momentum balance. The increase in the wind stress i s generally balanced by an increase in the barotropic pressure gradient corresponding to a surface setup, and bottom friction, particularly on the ocean side where the residual velocities are stronger Nonlinear advection during storms lost some strength due to the overwhelming wind stress An analysis of the evolution of eddies shows that the residual eddies inside the estuary is maintained by nonlinear advection and wind stress. The residual eddies are persistent and strong when the nonlinear advection is dominant. However the intensification of the relatively uniform wind stress during tropical cyclones inside the estuary causes the residual eddies to weaken or disrupt

PAGE 111

111 depending on the strength of the wind stress Baroclinic pressure gradient bec o me s significant after several storms due to the increased precipitation and river discharge inside the system especially in 2004, since the river discharge and precipitation were 3 times larger than those in 2008. Baroclinic pressure gradient gains strength generally inside the inlet due to the exchange of freshwater and high salinity ocean water. Other forcing functions in the momentum equations do not contribut e significantly to the balance. Coriolis, horizontal d iffusion and local acceleration a re almost zero. The major findings for this study can be summarized as: 4 residual eddies form around SAI during a typical tidal cycle caused by the nonlinear advection; Cross sectional residual flow is highly variable wit h location of cross section and dominant forcing s within the St. Augustine Inlet; Around St. Augustine Inlet, r esidual momentum balance during a typical tidal cycle is between barotropic pressure gradient and advection During tropical cyclones, uniform wi nd stress overwhelms the nonlinear advection inside the estuary to disrupt the residual eddies The duration of variation in residual flow structure and momentum balance depends on the strength and duration of tropical cyclones and can range from 34 days t o 12 days. During tropical cyclones, w ind speed and direction determines spatial structure of residual circulation at the S t. Augustine Inlet Baroclinic pressure gradient was less prominent in 2008 compared to 2004 simulation period, due to lower precipitation and river discharge

PAGE 112

112 Stratification inside the estuary is very small and r esidual flow rate at inlets play major role determining estuarine salinity Tidal and residual flow rate at S t. Augustine Inlet is 4 5 times larger than other 2 inlets

PAGE 113

113 REFERENCES Barron, C. N., A. B. Kara, R. C. Rhodes, C. Rowley, and L. F. Smedstad, 2007: Validation Test Report for the 1/8 Global Navy Coastal Ocean Model Nowcast/Forecast System. Boesch, D. F., R. J. Diaz, and R. W. Virnstein 1976: Effects of tropical storm Agnes on soft bottom macrobenthic communities of the James and York estuaries and the lower Chesapeake Bay. Chesapeake Science, 17 246 259. Cheng, R. T., and V. Casulli, 1982: On Lagrangian residual currents with applications in south San Francisco Bay, California. Water Resour. Res., 18 16521662. Davis, J. R., and Y. P. Sheng, 2003: Development of a parallel storm surge model. Int. J. Numer. Meth. F luids 42, 549 580. Fischer, H. B., 197 6: Mixing and dispersion in estuaries. Annu. Rev. Fluid Mech., 8 107133. Fischer, H. B., E. J. List, R. C. Y. Koh, J. Imberger, and N. H. Brooks, 1979: Mixing in inland and coastal waters. Academic Press, 483 pp. Garvine, R. W., 1985: A simple model of estuarine subtidal fluctuations forced by local and remote wind stress. J. Geophys. Res. 90 11945 11948. Gong, W., J. Shen, and W. G. Reay, 2007: The hydrodynamic response of the York River estuary to Tropical Cyclone Isabel, 2003. Estuar. Coast. Shelf S ci., 73 695710. Hench, J. L., and R. A. Luettich, 2003: Transient Tidal Circulation and Momentum Balances at a Shallow Inlet. J. Phys. Oceanogr., 33 913 932. Holland, G. J., 1980: An analytic model of the wind and pressure profiles in hurricanes. Mon. Weather Rev., 108 12121218. Janzen, C. D., and K. C Wong, 2002: Windforced dynamics at the estuary shelf interface of a large coastal plain estuary. J. Geophys. Res., 107 (C10) Jelesnianski, C. P., J. Chen, and W. A. Shaffer, 1992: SLOSH: sea, lake, and overland surges from hurricanes. Technical Report. NWS, 48, 71 pp. Kim, T., Y. P. Sheng, and K. Park, 2010: Modeling water quality and hypoxia dynamics in Upper Charlotte Harbor, Florida, U.S.A. during 2000. Estuar. Coast. Shelf Sci., 90 250 263 Kjerfve, B., 1978: Bathymetry as an Indicator of Net Circulation in Well Mixed Estuaries. Limnol. Oceanogr., 23, 816821.

PAGE 114

114 Li, C., and J. O'Donnell, 1997: Tidally driven residual circulat ion in shallow estuaries with lateral depth variation. J. Geophys. Res., 102 27915 27929. Li, C., and J. ODonnell, 2005: The Effect of Channel Length on the Residual Circulation in Tidally Dominated Channels. J. Phys. Oceanogr., 35 1826 1840. Li, C., 2006: Modeling of Bathymetry Locked Residual Eddies in Well Mixed Tidal Channels with Arbitrary Depth Variations. J. Phys. Oceanogr., 36 19741993. Li, M., L. Zhong, W. C. Boicourt, S. Zhang, and D. L. Zhang, 2006: Hurricaneinduced storm surges, currents and destratification in a semi enclosed bay. Geophys. Res. Lett., 33 Li, C., C. Chen, D. Guadagnoli, and I. Y. Georgiou, 2008: Geometry induced residual eddies in estuaries with curved channels: Observations and modeling studies. J. Geophys. Res., 113 C 01005. Liu, J. T., 1992: The influence of episodic weather events on tidal residual currents: A case study at Sebastian Inlet, Florida. Estuaries and Coasts, 15, 109 121. Lynch, M. P., 2005: An unprecedented scientific community response to an unprecedent ed event: Tropical Storm Agnes and the Chesapeake Bay. Hurricane Isabel in Perspective, Edgewater, MD., Chesapeake Research Consortium, 2935. Mallin, M. A., M. H. Posey, G. C. Shank, M. R. McIver, S. H. Ensign, and T. D. Alphin, 1999: Hurricane effects on water quality and benthos in the Cape Fear watershed: natural and anthropogenic impacts. Ecol. Appl., 9 350 362. Mukai, A. Y., J. J. Westerink, and R. A. Luettich, 2002: Guidelines for using the Eastcoast 2001 database of tidal constituents within the western North Atlantic Ocean, Gulf of Mexico and Caribbean Sea, Coastal and Hydraulics Technical Note CERD. Vicksburg, Virginia: US Army Corps of Engineers, ERDC/CHL CHETNIV 40, 20 Murty, T. S., F. G. Barber, and J. D. Taylor, 1980: Role of Advective Ter ms in Tidally Generated Residual Circulation. Limnol. Oceanogr., 25, 529 533. Owen, B., and White, S., 2005: National Estuarine Research Reserve System 10th anniversary report on the System wide Monitoring Program (SWMP) data applications, 19952005. Paerl, H. W. and Coauthors, 2001: Ecosystem impacts of three sequential hurricanes (Dennis, Floyd, and Irene) on the United States' largest lagoonal estuary, Pamlico Sound, NC. Proc. Natl. Acad. Sci. U. S. A., 98, 5655. Paramygin, V. A., 2009: Towards a r eal time 24/7 storm surge, inundation and 3D baroclinic circulation forecasting system for the state of Florida. Ph.D. dissertation, University of Florida,

PAGE 115

115 Pingree, R. D., and L. Maddock, 1977: Tidal eddies and coastal discharge. J. Mar. Biol. Ass. U. K. 57 869 875. Powell, M. D., S. H. Houston, L. R. Amat, and N. Morisseau Leroy, 1998: The HRD real time hurricane wind analysis system. J. Wind Eng. Ind. Aerodyn., 77 53 64. Powell, M., R. Thieke, and A. Mehta, 2006: Morphodynamic relationships for ebb and flood delta volumes at Florida's tidal entrances. Ocean Dynamics, 56, 3 4. Prandle, D., 1978: Residual Flows and Elevations in the Southern North Sea. Proceedings of the Royal Society of London.Series A, Mathematical and Physica l Sciences, 359 pp. 189 228. Sherman, B. (2003, 09/17/2003). Stage is set for Isabel to impact Chesapeake Bay ecosystem. Retrieved 10/25/2006, from http://www.publicaffai rs.noaa.gov/releases2003/sep03/noaa03r450.html Sheng, Y. P., 1987: On modeling threedimensional estuarine and marine hydrodynamics. Three dimensional models of marine and estuarine dynamics (Ed: J.C.J. Nihoul), Elsevier, 35 54. Sheng, Y. P., and C. Vill aret, 1989: Modeling the effect of suspended sediment stratification on bottom exchange processes. J. Geophys. Res., 94, 1442914444. Sheng, Y. P., 1990: Evolution of a ThreeDimensional Curvilinear Grid Hydrodynamic Model for Estuaries, Lakes, and Coastal Waters: CH3D. Estuarine and Coastal Modeling I, ASCE 40 49. Sheng, Y. P., V. A. Paramygin, V. Alymov, and J. R. Davis, 2005: A Real Time Forecasting System for Hurricane Induced Storm Surge and Coastal Flooding. Estuarine and Coastal Modeling IX, Charle ston, South Carolina, USA, ASCE 585602. Sheng, Y. P., B. Tutak, J. R. Davis, and V. Paramygin, 2008: Circulation and Flushing in the Lagoonal System of the Guana Tolomato Matanzas National Estuarine Research Reserve (GTMNERR), Florida. J. Coast. Res., 5 5 9 25. Sheng, Y. P., and T. Kim, 2009: Skill assessment of an integrated modeling system for shallow coastal and estuarine ecosystems. J. Mar. Syst., 76, 212 243. Sheng, Y. P., and V. Paramygin, 2010: Forecasting Storm Surge, Inundation, and 3D Circulat ion along the Florida Coast. Estuarine and Coastal Modeling XI ASCE Sheng, Y. P., V. Alymov, and V. A. Paramygin, 2010a : Simulation of storm surge, wave, currents, and inundation in the Outer Banks and Chesapeake Bay during Hurricane Isabel in 2003: The Importance of waves. J. Geophys. Res., 115 C04008

PAGE 116

116 Sheng, Y. P., Y. Zhang, and V. A. Paramygin, 2010b : Simulation of storm surge, wave, and coastal inundation in the Northeastern Gulf of Mexico region during Hurricane Ivan in 2004. Ocean Modelling, 35 314 331. Smith, N. P., 2001: Seasonal scale transport patterns in a multi inlet coastal lagoon. Estuar. Coast. Shelf Sci., 52 15 28. Tutak B. and Y. P. Sheng, 2010a : Simulating the Hydrodynamic Response of NCB Estuary to 2004 and 2008 Tropical Cyclones In pr eparation. Tutak B. and Y. P. Sheng, 2010b: Effect of Tropical Cyclones on Residual Circulation and Momentum Balance in a Subtropical Estuary and Inlet in 2004. In preparation. Tutak B., Y. P. Sheng, and V. Paramygin, 2010: Effect of Tropical Cyclones on R esidual Circulation and Momentum Balance in a Subtropical Estuary and Inlet Observation and Simulation. In preparation Tee, K. T., 1977: Tide Induced Residual Current --Verification of a Numerical Model. J. Phys. Oceanogr., 7 396 402. Valiela, I. and Coauthors, 1998: Ecological effects of major storms on coastal watersheds and coastal waters: Hurricane Bob on Cape Cod. J. Coast. Res., 14 218238. Valle Levinson, A., K. C. Wong, and K. T. Bosley, 2002: Response of the lower Chesapeake Bay to forcing from Hurricane Floyd. Cont. Shelf Res., 22, 17151729. Valle Levinson, A., C. Reyes, and R. Sanay, 2003: Effects of Bathymetry, Friction, and Rotation on Estuary Ocean Exchange. J. Phys. Oceanogr., 33 23752393. Walker, N. D., 2001: Tropi cal storm and hurricane wind effects on water level, salinity, and sediment transport in the river influenced AtchafalayaVermilion Bay system, Louisiana, USA. Estuaries and Coasts, 24 498 508. Wang, D., 1979: Subtidal Sea Level Variations in the Chesapeake Bay and Relations to Atmospheric Forcing. J. Phys. Oceanogr., 9 413 421. Waterhouse, A. F B. Tutak, A. Valle Levinson, and Y. P. Sheng, 2010: Response of residual flows to two tropical storms in a subtropical tidal inlet. In preparation. Webb, B. M., J N. King, B. Tutak, and A. Valle Levinson, 2007: Flow structure at a trifurcation near a North Florida inlet. Cont. Shelf Res., 27 15281547. W esterink, J. J. and Coauthors, 2008: A basinto channel scale unstructured grid hurricane storm surge model applied to southern Louisiana. Mon. Weather Rev., 136, 833 864.

PAGE 117

117 Winant, C. D., 2008: Three Dimensional Residual Tidal Circulation in an Elongated, Rotating Basin. J. Phys. Oceanogr., 38 12781295. Wong, K. C ., 1994: On the nature of transverse variability in a coastal plain estuary. J. Geophys. Res., 99 1420914222. Zimmerman, J., 1978: Dispersion by tideinduced residual current vortices. Elsevier Oceanography Series, 23, 207216.

PAGE 118

118 BIOGRAPHICAL SKETCH Bilge Tutak was born in 1980 in Bandirma, Turkey. He received his B.S. degree in Civil Engineering from Bogazici University in 2003 where he continued his M.S. degree until 2005. He studied wave mechanics and meshless methods for numerical models. He started his Ph.D. degree in August 2005 at University of Florida under the supervision of Y. Peter Sheng. He worked on estuarine and ocean modeling. He was a National Estuarine Research Reserve (NERR) Graduate Fellow for 3 years with a funding for his research on the effects of extreme events on estuaries. After obtaining his Ph.D. degree, he wishes to pursue an academic career.