Citation |

- Permanent Link:
- http://ufdc.ufl.edu/UF00091383/00001
## Material Information- Title:
- Efectiveness of fine sediment trapping in an estuary
- Series Title:
- Efectiveness of fine sediment trapping in an estuary
- Creator:
- Paramygin, Vladimir A.
- Place of Publication:
- Gainesville, Fla.
- Publisher:
- Coastal & Oceanographic Engineering Dept. of Civil & Coastal Engineering, University of Florida
- Language:
- English
## Record Information- Source Institution:
- University of Florida
- Holding Location:
- University of Florida
- Rights Management:
- All applicable rights reserved by the source institution and holding location.
## UFDC Membership |

Full Text |

UFL/COEL-2002/002
EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY by Vladimir A. Paramygin THESIS 2002 Coastal & Oceanographic Engineering Program Department of Civil & Coastal Engineering 433 Weil Hall *P.O. Box 116590 Gainesville, Florida 32611-6590 UNIVERSITY OF FLORIDA UFL/COEL-2002/002 EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY by Vladimir A. Paramygin THESIS 2002 EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY By VLADIMIR A. PARAMYGIN A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2002 ACKNOWLEDGMENT I wish to express my gratitude to Dr. Ashish J. Mehta, chainian of my graduate committee, for his guidance in my education and research, and for his patience, support and understanding. I also wish to thank the members of my graduate committee, Dr. Robert J. Thieke and Dr. Daniel M. Hanes, for their great teaching effort and assistance. Special thanks go to Dr. Earl J. Hayter for his assistance with my research effort, and Sidney L. Schofield for help in the laboratory and field work. I wish to thank my parents and sister for their love and trust, my many friends for their support, encouragement and help, and the faculty and staff of the Coastal and Oceanographic Engineering Program of the Department of Civil and Coastal Engineering for their help during my stay. Data used in this study are in part derived from field work conducted for the St. Johns River Water Management District (SJRWMD) of Palatka, Florida. Thanks are due to Dr. Chandy John of the SJRWMD for his technical assistance throughout this study. Finally, very special thanks go to my friends Oleg Mouraenko, Leonid Parshukov, Angela Chulkova, Yuri Melentyev and Ulyana Merkulova for their help and support. TABLE OF CONTENTS pne ACKN OW LED GM EN T ..................................................................................................... ii LIST O F TA BLES .............................................................................................................. v LIST OF FIGU RES .......................................................................... I ................................. vi EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY .................. viii I INTRODU CTION ......................................................................................................... I 1. 1 Problem Statem ent .................................................................................................. 1 1.2 Cedar River ............................................................................................................. 2 1.3 Objective and Tasks ................................................................................................ 3 1.4 Outline of Chapters ................................................................................................. 3 2 M ETHOD OF ANA LY SIS ........................................................................................... 5 2.1 Introduction ............................................................................................................. 5 2 .2 F lo w F ie ld ............................................................................................................... 5 2.3 Sedim ent Transport ............................................................................................... 10 2.4 Settling V elocity Calculation ................................................................................ 11 2.4.1 Background .................................................................................................. 11 2.4.2 Particle Density and Fractal Representation ................................................ 15 2.4.3 Settling Velocity ........................................................................................... 15 2.4.4 Floc Growth and Breakup Functions ........................................................... 17 3 CEDA R RIVER ESTUA RY ....................................................................................... 20 3.1 D escription of the Estuary ..................................................................................... 20 3.2 Tide, W aves, Current, Salinity and W ind Data ..................................................... 22 3 .2 .1 T id e d ata ....................................................................................................... 2 2 3.2.2 Current data .................................................................................................. 23 3.2.3 W ind data ..................................................................................................... 24 3.2.4 W ave data ..................................................................................................... 26 3.2.5 Salinity data .................................................................................................. 27 3.3 D ischarge D ata ...................................................................................................... 28 3.4 Sedim ent Concentration ........................................................................................ 29 3.5 Bed Sedim ent Distributions .................................................................................. 30 iii 4 ASSESSMENT OF SEDIMENT TRAPPING EFFICIENCY ................................... 37 4.1 Flow M odel Setup, Calibration and Validation .................................................... 37 4. 1.1 Cedar/Ortega/St. Johns Rivers M odel Setup ................................................ 37 4.1.2 Cedar River M odel Setup ............................................................................. 43 4.2 Sedim ent Transport M odel Setup and Calibration ................................................ 50 4.2.1 Sedim ent Transport M odel Setup ................................................................. 50 4.2.2 Bed Erosion .................................................................................................. 52 4.2.2 Settling Velocity and Deposition ................................................................. 53 4.3 Trapping Effi ciency Analysis ................................................................................ 57 4.3.1 Treatm ent Plan ............................................................................................. 57 4.3.2 Sedim ent Trap Setup .................................................................................... 57 4.3.3 Effect of Trap Efficiency on Settling Flux Dow nstream ............................. 58 5 CON CLU SION S ......................................................................................................... 61 5 .1 S u m m a ry ............................................................................................................... 6 1 5.2 Conclusions ........................................................................................................... 61 5.3 Recom m endations for Further W ork ..................................................................... 62 APPENDICES A WATER DISCHARGE ESTIMATION BASED ON ADCP ..................................... 63 B SETTLING VELOCITY AND FLOC SIZE CALCULATIONS .............................. 65 B I In tro d u ctio n .......................................................................................................... 6 5 B.2 Settling Velocity Calculations .............................................................................. 65 B.3 Particle Size Calculations ..................................................................................... 67 REFEREN CES .................................................................................................................. 69 BIOGRAPH ICA L SK ETCH ............................................................................................ 71 iv LIST OF TABLES Table Page 3-1: Wind speed/direction distribution. Dominant speed/directions are highlighted. 25 3-2: Cedar River cross-section discharges, May 17, 2001 ................................ 29 3-3: Sediment concentrations from water samples, May 17, 2001...................... 30 3-4: Statistical values associated with bed sediment distribution ....................... 32 4-1: TSS removal efficiencies of treatment systems in Florida (after Harper, 1997) 57 4-2: Comparison of sites with different removal efficiencies with a no-trapping scenario................................................................................. 59 4-3: Summary of the effect of treatment on sediment load in the confluence area ...60 B-i1: Data from settling column tests with Ems-Dollard mud............................ 68 LIST OF FIGURES Figure Page 2-1: Effect of sediment concentration and fluid shear stress on the median floc d iam eter (D y er, 19 89) .......................................................................................... 13 2-2: Relationship between settling velocity and floc diameter in still water, based on e q u atio n (2 .2 9 ) ..................................................................................................... 17 3-1: Cedar/Ortega Rivers estuary, (within the light rectangular area), aerial photo, M a y 19 9 8 .............................................................................................................. 2 0 3-2: Cedar/Ortega Rivers data collection and sediment treatment (Wet Detention S y stem ) sites I an d 2 ............................................................................................. 2 1 3-3: Tidal ranges at stations TG I -TG3. Cumulative frequency distribution based on record obtained during 09/29/00-10/18/01 ............................................................ 23 3-4: Current speed at station WGC (mouth of the Ortega in St. Johns River). Cumulative frequency distribution based on record obtained during 02/05/01 0 3 /0 8 /0 1 ................................................................................................................. 2 4 3-5: Wave spectrum based on 10/02/01 -25/04/01 record at WGC ............................. 26 3-6: Significant wave height based on spectral analysis ............................................. 27 3-7: Cumulative distribution of the significant wave height at the mouth of the Ortega R iver in the St. Johns R iver ................................................................................... 27 3-8: Cumulative distribution of salinity at stations TG I -TG3; 10/27/00-11/26/00 ...... 28 3-9: Cedar River cross-section discharge and Cedar/Ortega tide data, May 17, 2001. 29 3-10: Moisture content distribution (%). Based on 1998 sampling ................................ 33 3-11: Organic content distribution (%). Based on 1998 sampling ................................ 34 3-12: Solids content distribution (%). Based on 1998 sampling ................................... 35 3-13: Thickness of soft deposit in the study area based on core thicknesses in 1998 sam p lin g ............................................................................................................... 3 6 4-1: Areas covered by the two (coarse grid and fine grid) models .......................... 38 4-2: Cedar/Ortega/St. Johns River grid with open boundary locations .................... 39 4-3: Cedar/Ortega/St. Johns River bathymetry. Depths are in meters ..................... 40 4-4: Cedar/Ortega/St. Johns River open boundary condition (BC7) during 2001 showing water surface elevation time-series and cumulative distribution ..... 41 4-5: Cedar/Ortega/St. Johns River tributaries discharge, cumulative distribution ....... 42 4-6: C edar R iver m odel grid .................................................................................... 44 4-7: Cedar River bathymetry. Depths are in meters ................................................ 45 4-8: Measured and simulated water level variations at the downstream boundary of the C edar R iver ................................................................................................... . 4 7 4-9: Water surface elevation at three control points in Cedar River ....................... 48 4-10: Measured and simulated discharges through the Cedar River cross-section ........ 48 4-11: Measured and simulated discharges through the Ortega River cross-section (north cross-section of the confluence) ...................................................................... 49 4-12: Measured and simulated discharges through the Ortega River cross-section (south cross-section of the confluence) ...................................................................... 49 4-13: Depth averaged TSS concentrations at the Cedar River cross-section near the Cedar/Ortega confluence simulated by the coarse grid model (May 17, 2001).... 51 4-14: Depth averaged TSS concentrations from the water sample data, collected at the Cedar River cross-section near the Cedar/Ortega confluence (May 17, 2001)..... 51 4-15: Bed erosion rate function obtained from laboratory experiments on mud from the C edar/O rtega R ivers .......................................................................................... 53 4-16: Calculated floc size as a function of shear stress and concentration ................. 55 4-17: Settling velocity curve based on laboratory tests in a settling column using sediment from the Cedar River and vicinity .................................................... 56 B-l: Settling velocity calculation test results, and comparison with data of Wolanski et al. (1992) using sediment from Townsville Harbor, Australia ........................ 66 B-2: Floc growth with time measured and predicted for River Ems-Dollard mud ....... 68 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY Vladimir A. Paramygin May 2002 Chairman: Ashish J. Mehta Major Department: Civil and Coastal Engineering A study of the effectiveness of fine sediment trapping in the Cedar River estuary in north Florida is presented. A combined hydrodynamic and sediment transport model has been used to simulate discharge, water level variation and suspended sediment flux at various sections in the river. A settling velocity sub-model accounting for the effects of floc growth and breakup due to turbulence in the water column has been incorporated in the model. The simulated discharges and surface elevations at the downstream open boundary of Cedar River at its confluence with the Ortega River are found to agree reasonably well with measurements. It is shown that trapping sediment upstream in Cedar River would have only a minor effect on sediment deposition in the downstream region, where contaminated fine sediment has accumulated. This is so because the majority of the sediment is apparently derived from creeks entering the river within its middle reach. On the other hand, sediment entrapment closer to the confluence of the Cedar and Ortega rivers appears to be able to measurably reduce sediment transport to the confluence and, therefore, can be expected to lower the flux of sediment-bound contaminants out of Cedar River. CHAPTER 1 INTRODUCTION 1 .1 Problem Statement Sediment shoaling in estuarine environments can create significant problems such as degradation of water quality and concentration of organic matter and contaminants. Accumulated organic-rich sediment can increase contaminant loads in these waters, because contaminants such as polychlorinated biphenyls (PCBs) and polyaromatic hydrocarbons (PAHs) are preferentially bound to organics (National Research Council, 2001). A commonly implemented solution to reduce sedimentation is the construction of a sediment trap. Such traps can be of different types; however, all of them rely the same basic mechanism decreasing the speed of the flow, thus allowing the larger portion of the suspended sediment load to settle out in the trap. Traps can be on-line or off-line. One example of an on-line trap is a dredged trench along the submerged bottom, which reduces the flow velocity and causes the material to settle there. An off-line trap is made by artificially diverting part of the flow into a natural/artificial pond, which would reduce flow velocity and increase deposition. Selecting the Cedar River in northern Florida as a case study, Stoddard (2001) examined the efficiency of a trap trenched at the bottom of the river. In the present study, the efficiency of a trap created by ponding along the side of the same river is explored. For the present purposes, trap efficiency will be determined by the sediment removal ratio, i.e., the percentage by which the effluent sediment load is reduced with 2 respect to the influent load (Ganju, 200 1). By creating efficient traps, much of the detrimental effects of excess sediment and unwanted contaminants entering the system can be curtailed. 1.2 Cedar River Cedar River estuary occurs in northeast Florida. Trapping contaminants in this river has become essential due to elevated concentrations of sediment-bound PCBs in water resulting from leaching of sediment and runoff from the site of a chemical company since January 1984. The site is located approximately 0.5 kmn east of the Cedar River near its headwaters, adjacent to municipal stor drains and drainage ditches. There, fire destroyed several tanks storing high concentrations (4,425 ppm) of PCB-laden oils and other materials. It is believed that a combination of damage to the storage tanks and the fire-fighting effort caused PCBs to enter the Cedar River basin. The surrounding groundwater and soils were sampled extensively in 1989, and the concentrations were still significantly above the regulated amount of 50 ppm. Estuaries characteristically trap significant quantities of particulate matter through a wide variety of physical and biochemical processes. Fine-grained sediments play an important role in these processes. Due to relatively strong currents, fine sediments, which are mixtures of clay- and silt-sized material, are usually very mobile. In the Cedar River they are admixed with organic matter derived from local terrestrial and aquatic sources. Fine sediment transport is mainly defined by the hydrodynamic action, which advects the suspended matter and provides the bed erosion force. Also, turbulence plays a major role in the flocculation of fine, cohesive sediments. Flocs are formed by the joining of individual particles and can greatly affect the settling velocity of particulate matter. The St. Johns River Water Management District of Palatka, Florida is considering the possibility of establishing off-line sediment traps upstream along the Cedar River, and would like to have an estimate of the influence of this entrapment on sediment influx at the downstream end of Cedar River, at its confluence with the Ortega River, where heavy accumulation of PCB-laden, organic-rich sediment has occurred. 1.3 Objective and Tasks The objective of this study was to determine the effect of traps with different efficiencies at selected locations upstream in Cedar River to sediment flux at the confluence of Cedar and Ortega Rivers downstream. Several tasks were undertaken to achieve this objective including: I) Use of data to characterize the nature of flow. Data included tidal elevations, current velocities, wind speed and direction, salinity, streamfiows at the major tributaries of the Cedar River and the Cedar River itself, and suspended sediment concentrations. 2) Modeling the flow field using a numerical code to determine water velocities, water surface elevations and bed shear stresses. 3) Modeling fine sediment settling velocity as a function of the local flow conditions. 4) Use of a sediment transport model (with implemented settling velocity model) to determine suspended sediment concentrations within the modeled domain. 5) Using the calibrated flow and sediment models, modeling flow and sediment transport in the estuary with the sediment traps (with three assumed efficiencies 30%, 60% and 90%) and without the traps. 6) Comparison of the results of above modeling in terms of sediment transport at the downstream end of Cedar River to assess the effects of above traps. 1.4 Outline of Chapters Chapter 2 describes the flow and sediment transport model used to evaluate trap efficiency. Chapter 3 contains the field data collected for this study. Chapter 4 describes 4 the calibration and validation of the model using measured data. Finally, Chapter 5 contains the summary and conclusions, followed by bibliography. CHAPTER 2 METHOD OF ANALYSIS 2.1 Introduction This chapter provides a description of the hydrodynamic and sediment models that were used to model flow and the sediment transport in the Cedar River. The chapter gives basic equations, numerical method used to solve these equations, and the capabilities and limitations of the models for problem analysis. The Environmental Fluid Dynamics Code (EFDC) used herein implements a numerical algorithm for estuarine flows (Hamrick, 1996). It contains a three-dimensional, hydrostatic flow model, as well as a compatible sediment model. 2.2 Flow Field The coordinate system of the mi-odel is curvilinear and orthogonal in the horizontal (x,y) plane. In the vertical, z direction, which is aligned with the gravity vector, it is stretched to follow the bottom topography and free surface displacement (G-grid). A level 2.5 turbulence closure scheme (Mellor and Yamada, 1982) in the hydrodynamic model relates the turbulent viscosity and diffusivity to the turbulence intensity and a turbulence length scale. An equation of state relates density to pressure, salinity, temperature and suspended sediment concentration (Hamrick, 1992). The momentum equations in the model are a, (i,iniHu)+ a, (in, Hu u)+ a, (m,Hvu)+ D (17m,m, wuj-I ,,) n ,Hiv SH, (p + p,,,,, + )+n,(,xz, + za,H)a p+ m,m. ,-D zu (2.1) H +1 j,-HAIu2 +Y K HAH,,U -mnCpDu- +v2 u a,(n,n, Hv) + a, (m,, Huv) + a,,( Hyv1v)+ a (,7m,,w, v)-f mim,,Hu =-m,H-, (p + p,,,,,, +)+nm,(a,.z +za,H)ap+a: ,.m a v (2.2) H nI m +8,. HAf,i +,, "xHA,, vV m.xmyc,,Du-+v2 v m,+m,, f m,m,,f u ,,m, + ,mn,. (2.3) (r,A, )= A,H (u, v) (2.4) where u and v are the horizontal velocity components in the x, y coordinate directions, respectively, and w, is the vertical velocity; inm, and inm,. are the scale factors of the horizontal coordinates; z and z, are the vertical coordinates of the free surface and bottom bed, respectively; H is a total water column depth; 5 = gz is a free surface potential; f, is the effective Coriolis acceleration and incorporates curvature acceleration terms with the Coriolis parameter, f as in equation (2.3); AH and A, are the horizontal and vertical turbulent viscosities, respectively, where A, relates the shear stresses to the vertical shear of the horizontal velocity components; the last terms in equations (2.1) and (2.2) represent vegetation resistance, where c, is a resistance coefficient and D, is the dimensionless projected vegetation area normal to the flow per unit horizontal area; and p,,,,, is the kinematic atmospheric pressure referenced to water density. The excess hydrostatic pressure in the water column is a p = -gHb = -gH(p Po)Po' (2.5) where p = p(p,S,T) (2.6) and p and P0 are the actual and reference water densities, respectively, and b is buoyancy. The three-dimensional continuity equation in the model is ,(,.,H)+ a,(,,m Hu)+ .(m,Hy)+ (mm,,, 1w)= Q, (2.7) and the corresponding vertically-integrated form of the continuity equation is (m,, H)+ (,, Hii)+ ,, (mIN ) Q (2.8) where Q,, represents volume sources and sinks including rainfall, evaporation, infiltration and lateral inflows and outflows having negligible momentum fluxes. Transport equations for temperature and salinity are a,(mHS)+ a, (mHuS)+ a,,(mHvS)+ a (nimwS)= a (mH 'A,,..S)+Q, (2.9) a, (mnHT)+ a, (mHuT)+ a, (mniHvT)+ a (mnT)= a (nmH A,a T)+ Q, (2.10) where Q, and Q, are source and sink terms, respectively, which include sub-grid scale horizontal diffusion and thenrmal sources and sinks, and Ah is the vertical turbulent diffusivity. Two transport equations determine the turbulent intensity and turbulent length scale as follows: a, (mxm Hq2 )+a, nHuq 2)+, (,Hvq )+D (m ,nwq2) =.ll mm Aq a q 2 2min Hq3 (2.11) H Z.B1I A, (az 1 2 V2 + (12+V2) + 2nm,n, ( u)2 + ( v)2) 7+ ,,c,,D,, (, + +v2) +gK,.b + Q, a, (nm,m,iHql21l) + (1 Huq2) a, (n, Hvq21)+ m.y. qnm,,wq- 1) =a 1m,m A,,q a_1 -!n H +E 1 +E 1 (2.12) H -BI kHz kH(1- Z) +min,,E, H(u) +(v)+ gK,,ab b+ q,,c,,D,,(u2 + v +, where (E,,E, E3) =(1.8,1.33,0.25) are empirical constants, Qq and Q, represent additional source-sink tenms, and the third term in the last line of both equations represents net turbulent energy production by vegetation drag with a production efficiency factor of q,,. Equation (2.4), which specifies the kinematic shear stress at the bed and free surfaces, provides the vertical boundary conditions for the solution of the momentum equations. At the free surface, the shear stress boundary conditions are given by the water surface wind stress (r,,r,)=( ,,rS,.)= c U, + V, (U,, V,,.) (2.13) where U,, and V,,. are the x and y components of the wind velocity, respectively, 10 m above the water surface. The wind stress coefficient for the wind velocity components is c = 0.001 P" (0.8 + 0.065 U2 + V ) (2.14) P"ii where PO and p'' are the air and water densities. At the bed, the shear stress components are considered to be related to the near-bed or bottom layer velocity components as follows: , ._)=( r,,,, )= c, + v7 (u ,,v1 ) (2.15) where "I and v1 are the bottom layer velocity components and the bottom stress coefficient is c1, n (2.16) \ .2zo which assumes that the near-bottom velocity profile is logarithmic. In equation (2.16) /C is the von Kannan constant, A, is the dimensionless thickness of the bottom layer, and zo z0 = is the dimensionless roughness height. H The vertical boundary conditions for the turbulent kinetic energy and length scale equations are 2 q : z=l (2.17) 2 q 2 =B3Tzh z=0 (2.18) 1=0 : z=0,l (2.19) The above set of equations forms a closed system that is solved by a numerical method (Hamrick, 1992). The model uses the finite volume method to bring the partial differential equation into a discrete form. The Smolarkiewicz (1983) scheme is used to solve for the 2D advection problem. An external/internal mode splitting procedure is implemented to increase the numerical efficiency of the code. 2.3 Sediment Transport The transport equation for a dissolved or suspended material having a mass per unit volume concentration C is a, (mnn,,HC)+ a,(m,HuC)+ (mHvC)+ a (mm,.wC)- a: (mn,,w, C) m, HK HK K,. +Q (2.20) =8,| HKa,C +a,. HKa.v +a m.,m.,-aC +Q. 1, m L 1,- H where KH and K, are the horizontal and vertical turbulent diffusion coefficients, respectively, w, is (a positive) settling velocity when C represents suspended matter, and Q, represents external, and reactive internal, sources and sinks. Due to a small numerical diffusion that remains inherent in the scheme used to solve the sediment transport equation, the horizontal diffusion terms are omitted from equation (2.20). This results in I,(mX. HS, )+ ,(m ,HvS,) + .(mHvS,)+ a.(nn,wS,) I IK 1 (2.21) -a Z(mws,)= ,, ,, aS, +Q, +Q, H where S, represents the concentration of the j -th sediment class. Source-sink are represented by two terms: an external part, which would include point and non-point source loads, and an internal part, which could include reactive decay of organic sediment, or exchange of mass between sediment classes when floe growth and breakup are simulated. The vertical boundary conditions for equation (2.2 1) are -K 1 ZSI 14S = O : Z = H (2.22) where J. is the net water column-bed exchange flux (Hamrick, 1992). 2.4 Settling Velocity Calculation 2.4.1 Background A settling velocity algorithm was implemented, as part of the present study, in the sediment transport code. The algorithm calculates the settling velocity of the particles by accounting for the floc growth and breakup processes that occur for fine-grained sediment in estuarine and coastal waters due to different mechanisms. As a result, instead of using the settling velocity measured in a laboratory settling column in still water directly, the model is merely calibrated using laboratory data. There are a number of models in which the settling velocity is expressed as an analytical function of the shear rate and the sediment concentration. Also, there are some models that take a different approach in which the settling velocity only depends on the properties of the primary particles. Generally, the former use the median (or mean) size of the particles and the latter use a multi-class approach, in which particle sizes are defined by a discrete distribution function. Both approaches have their advantages and disadvantages. The former models usually depend on empirical coefficients, about which a very little is known in most cases. The latter, however, require significant computational time and, even though they are useful in simulating simple cases, high memory and processing power requirements make them almost unusable for simulations with large grids and for significant periods of time. Accordingly, in this study an attempt has been made to implement a simple settling velocity model dependent on empirical coefficients, but which includes the basic physics of floc growth and breakup. Floc growth/breakup can be triggered by different mechanisms Brownian motion, turbulent motion and differential settling. The following model only accounts for the turbulence effect on growth/breakup. By analyzing theoretical results presented in the literature and carrying out their own experiments in a settling column, Stolzenbach and Elimelich (1994) concluded that the effect of the differential settling is minor and may be practically absent in many turbulent flow situations. This would be the case because the probability of the event when a particle with a large settling velocity overtakes a particle with a smaller settling velocity is small, due to the fact that the trajectories of the two particles tend to be deflected from one another. Also, the Brownian motion effect on growth/breakup can be considered to be negligible in estuaries (Winterwerp, 1998). Dyer (1989) presented a schematic description of the dependence of floc (median) diameter on both turbulence and sediment concentration, as shown in Figure 2- 1. FLOC DIAM Prn .y112 10-3 Figure 2-1: Effect of sediment concentration and fluid shear stress on the median floc diameter (Dyer, 1989). At very low concentrations and shear stresses, collisions are rare and the floc growth rate is very small. Increasing fluid shear increases the number of collisions, thus forming larger particles. A further increase in shear stress, however, causes the floc breakup process to dominate over the floc growth, thus decreasing floc size. Also, increasing sediment concentration increases the particle collision frequency, thus causing the median floc size to increase. However, above approximately 10 kg/m3, flocs start to disaggregate due to the low shear strength of the larger particles, which makes them fragile. A study of growthlbreakup of fine sediment based on multi-class grain size distribution, two- and three-body collisions, Brownian motion and differential settling was done by McAnally and Mehta (2001). They present a set of equations that characterize the growth/breakup mechanisms. These equations depend on primary particle properties, flow properties and two empirical parameters: a collision diameter function and collision efficiency. (However, collision efficiency can also be related to the collision diameter, thus the system only contains one empirical, heuristic parameter). They propose expressions for both, collision diameter and collision efficiency functions, based on dimensionless arguments, and provide fitted forls of those functions using experimental data: Collision diameter function: < p,,q;, + S T'E0' CE PF =H [~ JFIIJ ~ 4~J (2.23) Collision efficiency function: -(D / (-0.875 CEC "yDso) T, CEC0 ,,, ,[iP1 ) uSok (2.24) Ap,D3', (v +DI where R. is some function of the bracketed non-dimensional terms; indices i and in denote the colliding particle size classes; D = diameter of primary grain; D, = reference particle size; S = salinity; SO = reference salinity; T = fluid temperature, deg Celsius; T = reference temperature; D,, D,, = particle diameters, Ap, and Ap,, differential densities of the particles; v = kinematic viscosity of the fluid; u, = velocity of particles; r,, = maximum flow induced shear stress in a spherical particle; T.... = shear stress experienced by the m aggregate and CEC, CEC, = actual and reference cation exchange capacities, respectively. Parshukov (2002) presents a multi-class grain size settling velocity model, which implements the collision mechanisms presented in McAnally and Mehta (2001), and combines them with other turbulence related parameters and settling velocity expressions to represent growth/breakup. Testing of the model was done using laboratory data involving grid-generated turbulence and its effect on the settling of flocculated clays. 2.4.2 Particle Density and Fractal Representation For estuarine flocs the relation between the volumetric and mass concentrations, 0 and c, respectively, and between 0 and the number of particles per unit volume, n, is given by: 0K ;z =J.kP C=fnD' (2.25) where p, is the sediment density; p, floc density; p,, = water density; c = sediment concentration; f = is a shape factor (for spherical particles J = )r/6); and D = particle diameter. It is has been shown elsewhere that mud flocs can be treated as fractal entities (Krone, 1984, Huang, 1994). Kranenburg (1994) shows that the differential density Ap, can be related to the fractal dimension of the particles using the formulation SD 3-,il Ap,= p/ -p ,-(p, -p,.) D (2.26) where D is the particle diameter, Dp is the diameter of the primary particles and nf is the fractal dimension. The fractal dimension for strong estuarine flocs is found to be in a range of 2.1-2.3 (Winterwerp, 1998). 2.4.3 Settling Velocity A settling velocity function including the effects of both concentration and fluid shear rate was proposed by van Leussen (1994): (l+aG) (2.27) W., 0 (1 + bG2) where W'' and W,0 are the actual settling velocity and reference settling velocity, respectively, G is the dissipation parameter or the rate of flow shear, and a,b are empirical constants. Teeter (2001) proposed a more advanced functional relationship: W =a (I I 1 +-aG Jexp-a4 ) (2.28) C,,, L I +a3G- 2 1(.8 where C, and C,, are mass-weighted average upper and lower reference concentrations, respectively; and n, a,, a, a3 and a4 are the empirical parameters. Using the force balance for a settling particle, one can obtain an implicit formula for the settling velocity in still water, which depends on the fractal dimension (Winterwerp, 1998): w a (p, -p,.)g D3_,/ D 067-2 188/ p 1 + 0.15 Re6'7 (2.29) where a, #1 are the coefficient that depend on particle sphericity, these coefficients will be taken as 1 (spherical particles) here; li = dynamic viscosity and Re = settling Reynolds number, p,,WD1 /I Figure 2-2 shows the relationship between the floc diameter and the settling velocity in still water, as described by equation (2.29). D vs Ws 10 ...... 1 0 3 ; . . . . . . ... . . . . . . : .. . . . . . 102 nf=2 101 . .nf-2'3:":: :: i E .......... 2 E . . .. . 100 .... ....... .... ... ... .. ... ...... . 10-2. 101 102 10 104 D, mu Figure 2-2: Relationship between settling velocity and floc diameter in still water, based on equation (2.29). 2.4.4 Floc Growth and Breakup Functions In the following equations the effect of turbulence is expressed through the energy dissipation parameter G, defined as G dz (2.30) V ,LI where e = energy dissipation rate of flow, v = kinematic viscosity of the fluid, I dynamic viscosity of the fluid, = fluid shear stress and u = mean flow velocity. Levich (1962) determined the rate of coagulation of particles in a turbulent fluid by integrating the diffusion equation over a finite volume n: dn 3 =--e,.e,GD3n2 (2.31) dt 2 where e.,e,, are the collision and diffusion efficiency parameters, respectively. Combining it with equations (2.25) and (2.26) yields the expression for the rate of growth of particle: dD_ 3 e et c GD"' D4 f = k, cGD4-, dt 2 f, nf p, (2.32) where 3exe D,"I- DI k 3 -k k (2.33) 2 f, nf p, nf p, The rate decay of particle due to breakup is suggested as: dD----"DG = I 1 i -kl,,G'1"' (D -D,, D '"'l (2.34) dt nf D, F D /2 k- aeD- D 1, (2.35) nf F, nf F, where p and q are the empirical parameters; eh is a floc breakup efficiency parameter and F, is the floc strength (assumed to remain constant, due to the fractal structure of flocs and determined by the number of bonds in a plane of failure) (Winterwerp, 1998). Combining the rates due to floc growth and floc breakup yields the expression for the net rate of change of rate due to turbulence: dD k,, c GD"' -3D4-if ,/, G'"+D-I'D'"+(D-D (2.36) dt nf p, 1 nf F Winterwerp (1998) assumed that the parameters k and k' remain constant, noting that the empirical coefficients they depend on are poorly known. In the present study, k, is assumed to remain constant, since it depends only on floc strength (which as noted remains unchanged due to the fractal nature of the particle). However, since the growth process is more complicated than breakup, we will allow k" to remain a variable, enabling a differentiation between the flocculation and hindered modes of floc settling during the calibration process. In the foner mode the settling velocity increases with increasing concentration, whereas in the latter mode, which occurs at higher concentrations, the settling velocity decreases with increasing concentration. Note, however, that limitations are imposed on the choice of k', based on the reported ranges of e, = O{ 10-2} and e,, 0.5-1.0 (Levich, 1962; O'Melia, 1985). Winterwerp (1998) estimated k = O{10-5}. CHAPTER 3 CEDAR RIVER ESTUARY 3.1 Description of the Estuary The Cedar River estuary is contained within Duval County in northeast Florida Figure 3-1. Figure 3-1: Cedar/Ortega Rivers estuary, (within the light rectangular area), aerial photo, May, 1998. Both the Cedar River, and the Ortega River into which it flows, together empty into St. Johns River, which is connected to the Atlantic Ocean. Two main tributaries feed Cedar River--Butcher Pen Creek and Williamson Creek--along with several smaller tributaries. Another stream, Fishweir Creek also flow into the Ortega (Figure 3-2). B Site 1 Site 2 St John River WGC 0 o TG1 CO .. 14 1^04 TG2 0 Ortega River ADCP Transects *L 0 Ortega River 1 kilometer Figure 3-2: Cedar/Ortega Rivers data collection and sediment treatment (Wet Detention System) sites 1 and 2. Depths within the Cedar/Ortega system vary from 0.5 to 3 m with an average depth of just over 1 1n. Depths at the Cedar River vary from 0.3 in to 1.5 mn with an average of 0.5 m. 3.2 Tide, Waves, Current, Salinity and Wind Data 3.2.1 Tide data Three ultrasonic recorders (Infinities USA Inc.) were installed to measure tidal elevations; data included here are for the period of November 29, 2000 through May 17, 2001. The locations were chosen so as to cover a relatively large area of the estuary, and also to facilitate gauge installation/removal and data retrieval. All gauges were placed against bridge piers. Gauge locations are shown in Figure 3-2 (TGl-TG3). Tidal elevations were measured relative to the National Geodetic Vertical Datum (NGVD). Sampling interval was set at 30 min. From the ogive curves presented in Figure 3-3 we note that the median range was around 30 cm. Station TG2, being the closest to the St. Johns River, in the wide portion of the Ortega River, responded to the tide the most. TG1, in Cedar River responded significantly less than the other two, possibly due to fresh water outflows from Butcher and Fishweir Creeks, which would have opposed the tide. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Figure 3-3: Tidal ranges at stations TG1-TG3. Cumulative frequency distribution based on record obtained during 09/29/00-10/18/0 1. 3.2.2 Current data A tethered current meter (Endeco 174) was installed at location WGC (Figure 32). The data reported here is from an approximately 1 month-long record (speed and direction) collected from February 5 through March 8, 2001, until the device malfunctioned. Data sampling interval was 15 minutes. As seen from Figure 3-4, the current speed was below 30 cm/s 98% percent of the time, and below 25 cmls 95% of the time. Hence currents in the estuary are not very strong, and cannot be expected to result a high level of sediment transport under normal weather conditions. Current speed, cumulative distribution 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 Speed, rn/s Figure 3-4: Current speed at station WGC (mouth of the Ortega in St. Johns River). Cumulative frequency distribution based on record obtained during 02/05/01-03/08/01. 3.2.3 Wind data Wind record was obtained from the Jacksonville Naval Air Station for the period January 1, 1995 through December 31, 1998 (sampling was every 3 hours). Wind statistics derived from this record (Table 3-1) provide infonniation on the dominant wind speeds and directions. These data indicate a potentially complex dependence of wind on wind-driven currents in the estuary, especially because portions of the waterway reaches are lined by trees, while others have been cleared and developed. 6.6 1 3.2 0.0 1 0.0 15 171 382 486 332 174 78 32 10 4 1 1 1 4.9 21 184 389 413 287 129 82 32 19 9 4 1 3 4.6 47 413 722 778 627 343 236 154 109 59 17 8 2 4 10.3 43 328 579 575 402 211 135 96 53 24 5 5 5 1 1 7.2 25 164 310 362 288 137 122 49 22 3 4 4.3 23 213 419 461 373 213 146 111 31 7 3 4 5.8 23 148 298 377 321 181 145 97 27 10 4 1 1 4.8 1368 916 836 766 542 318 300 112 40 15 14 2 2 15.2 30 178 458 453 380 243 145 89 31 13 13 10 1 5 6.0 24 162 371 414 412 273 193 109 55 16 8 2 1 5.9 29 272 527 513 519 283 183 55 9 6.7 24 195 380 354 308 150 104 20 8 3 8 3 1 4.5 33 257 576 617 463 252 186 51 16 12 8 3 2 7.2 11 139 330 403 308 129 69 25 7 2 5 2 4.2 27 214 548 814 687 274 142 50 15 5 4 1 1 8.1 Table 3-1: Wind speed/direction distribution. Dominant speed/directions are highlighted. 2 3 4 5 6 7 8 9 10 11 12 13 14 15 % 0 1 5.1 Speed, M/S 24 48 72 96 120 144 168 70 zi .2 192 216 240 264 288 312 336 360 occurrence 11.5 120.8 1 22.7 1 18.2 1 9.6 3.2.4 Wave data A pressure transducer (Transmetrics Inc.) at WGC was used to obtain wave data. The sampling interval was 6 hours. Data included here are for the February 10 through April 25, 2001. The modal period was found to be 2 s. Figure 3-5 is a spectral representation of data. The variation of the significant wave height H,,,, corresponding to the modal period is shown in Figure 3-6. The cumulative distribution of H,,0, shown in Figure 3-7, indicates that it did not exceed -0.2 mn. This in turn implies a mild wave climate, due to the limited wind fetches in the St. Johns River. Wave action in the Cedar River is believed to be even milder, and is unlikely to contribute much to sediment transport except under severe conditions when comparatively large waves may break along the banks. 0.16, 0.14 0.12 0.1 E 0.08 0.06 "~0.060.02 0 160 140 00.5 Time, Days 400Frequency, Hz Figure 3-5: Wave spectrum based on 10/02/01 -25/04/01 record at WGC. 03 0.25 0.2 E 0.15 I 0.1 0 40 50 60 70 80 90 100 110 120 130 140 150 Time, days Figure 3-6: Significant wave height based on spectral analysis. Wave data. HmO, from spectral analysis. Cumulative distribution. 10 go 80 70 60 E 50 40 30 20 10 0 0.05 0.1 0.15 0 2 0.25 0 3 0.35 HmO, m Figure 3-7: Cumulative distribution of the significant wave height at the mouth of the Ortega River in the St. Johns River. 3.2.5 Salinity data Salinity data were collected at stations TG1-TG3 during October27November 26, 2000. Figure 3-8 shows the cumulative distribution of these data for each station. The water is generally brackish; at TG 1 in Cedar River the salinity is low due to freshwater outflows from the river itself as well as the creeks that flow into it. Salinity. Cedar/Ortega Rivers. Cumulative distribution. 100 - -TG1 TG2 90 II TG3 70 ' 60 I 40 30 20 10 1 0 2 4 6 8 10 12 14 Salinity, ppt Figure 3-8: Cumulative distribution of salinity at stations TG I-TG3; 10/27/00-11/26/00. 3.3 Discharge Data Discharge measurements were obtained using an ADCP (Acoustic Doppler Current Profiler) Workhorse 1200 kHz (RD Instruments Inc.), on May 17, 2001. Four cross-sections of the Cedar/Ortega River confluence area were selected for data collection for almost a full semi-diurnal tidal cycle. In addition to the flow data, water samples were collected for determination of sediment concentration. Due to the shallow nature of the estuary and poor perfonnance of ADCP in shallow waters, the data were found to have a somewhat qualitative significance. A large fraction of the total discharge had to be estimated, Appendix A describes the estimation algorithm, which closely follows the estimation implemented in WinRiver (RD Instruments software for ADCP systems). Table 3-2 presents the analyzed data for the Cedar River cross-section (all transects were made at the cross-section of the Cedar River near the confluence; see Figure 3-2). Positive discharge is directed west, and negative is directed east. Table 3-2: Cedar River cross-section discharges, May 17, 2001. ADCP transect # Time Discharge, rn3/s 6 12:18 +90 14 13:25 -23 24 15:46 -69 32 17:13 -14 40 19:31 -3 50 20:29 +1 The mean depth at the cross-section was 1.0 m and maximum depth 1.4 m. Figure 3-9 shows discharge plotted on the same scale as tide data at TG2. It can be seen from the figure that as expected, the discharge curve precedes the water elevation curve. Discharge data. Cedar River cross-section. 100 1 1 1 1 1 14 15 16 17 18 Time, hr Water level at Main Bridge Cedar/Ortega River 19 20 21 Figure 3-9: Cedar River cross-section discharge and Cedar/Ortega tide data, May 17, 2001. 3.4 Sediment Concentration As noted, during the ADCP survey on May 17, 2001, a set of water samples was also collected. All samples were filtered, dried and weighed to determine the sediment i-f) - concentrations. Table 3-3 gives the data thus obtained. The concentrations are characteristically low, and range between 8 and 10 1 mg/I. Table 3-3: Sediment concentrations from water samples, May 17, 2001 Sample # Concentration, Sape Concentration, Sample # Concentration, nmg/I Sample # mg m 114 31 14 61 16 2 17 32 14 62 15 3 16 33 13 63 16 4 17 34 17 64 20 5 8 35 18 65 15 6 22 36 17 66 15 7 14 37 17 67 14 8 13 38 15 68 16 9 35 39 11 69 16 10 23 40 11 70 26 11 15 41 15 71 15 12 15 42 14 72 14 13 20 43 18 73 101 14 15 44 12 74 19 15 14 45 19 75 17 16 15 46 37 76 16 17 19 47 13 77 15 18 16 48 17 78 57 19 13 49 16 20 27 50 19 21 33 51 17 22 17 52 21 23 19 53 16 24 16 54 18 25 14 55 13 26 16 56 13 27 17 57 16 28 17 58 15 29 16 59 16 30 13 60 16 3.5 Bed Sediment Distributions In order to represent bed sediment distribution patterns, bed-sampling data supplied by the St. Johns River Water Management District were used to generate the mi-aps showing the distribution of solids content, moisture content and organic content in the Cedar/Ortega Rivers. This set of maps, based on data obtained during March 3 October 2, 1998, were generated using approximation methods as follows. First, maps were generated using Matlab routines for surface fitting (meshgrid, griddata). These routines generate a rectangular grid covering the data set supplied to them. Grid values are then approximated by fitting the surface to the data points and detenrmining the values at the grid points. This approximation caused two problems: * Due to the lack of adequate spatial coverage of data, values at the river boundaries were automatically assigned zero values; * River boundary presence was not considered when generating the surface. In order to avoid these problems, first the boundary points were approximated. The method used for this approximation was as follows. The distance from each data point to each boundary point was calculated (the shortest distance possible following the river). Boundary points (V)were then evaluated as: e (3-2) le ' /-I in which v, data point value, a standard deviation of the dataset, and distance to data point i from boundary point j After generating the boundary points they were merged with the measured data points, and surface fitting functions were applied to the combined data. In this way contour maps were produced for moisture content (Figure 3 -10), organic content (Figure 3-1 1), solids content (Figure 3-12) and thickness of the deposit (Figure 3-13). Measurement points are displayed as black dots and have values besides them. Also, equal-percent contours are drawn to identify the areas where the percentage of moisture/organics/solids is approximately the same. Table 3-4 contains relevant statistics: minimum, maximum and the mean values of moisture content, organic content and solids content. We observe that the upstream reach of the Ortega River is characterized by the high organic content (30-35%), whereas the upstream reach of the Cedar River has a high solids percentage (25-30%). These trends reflect the more natural, vegetated surroundings of the Ortega versus more developed reaches of the Cedar. A thickness of the deposition layer is more or less constant in the Ortega/St. Johns River and is much smaller in the Cedar River. Table 3-4: Statistical values associated with bed sediment distribution Statistic Moisture content Organic content Solids content Minimum 54 6 16 Maxim-umn 84 51 46 Mean 76 21 24 Figure 3-10: Moisture content distribution (%). Based on 1998 sampling. Figure 3-11: Organic content distribution (%). Based on 1998 sampling. Figure 3-12: Solids content distribution (%s). Based on 1998 sampling. Figure 3-13: Thickness of soft deposit in the study area based on core thicknesses in 1998 sampling. CHAPTER 4 ASSESSMENT OF SEDIMENT TRAPPING EFFICIENCY 4.1 Flow Model Setup, Calibration and Validation Model setup was carried out in two steps. The area of interest lies within the Cedar River, while some of the data for calibration were available for various sites within the much larger Cedar/Ortega/St. Johns River estuary. It was therefore decided to calibrate and run the model with a coarse grid covering the Cedar/Ortega/St. Johns River estuary. Running the model in this way generated the downstream boundary conditions for the Cedar River model, which was then run for sediment trapping efficiency assessment. This procedure allowed for the use of a finer grid in the Cedar River, without significantly increasing the simulation time. Figure 4-1 shows the areas covered by the two model setups. Both, the coarse and fine grids are Cartesian because, as noted, the sediment transport model was found not conserve mass when run with a curvilinear grid. 4.1.1 Cedar/Ortega/St. Johns Rivers Model Setup The Cedar/Ortega/St. Johns River (Cartesian) grid and has dimensions of 160 by 300 cells, with a cell size of 50 by 50 mu (Figure 4-2). The boundary conditions are labeled BC 1 -BC7. The grid was o -stretched in the vertical direction with six horizontal layers. Figure 4-3 shows the bathymetry of the modeled domain. The Cedar/Ortega River portion of the domain is typically shallow with 1.5 2 mn depth in the channel thalweg, and an average depth of -1 m. Area covered by the Cedar River / grid only Area covered by the Cedar/Ortega/St. Johns grid only / /f Area covered by both grids Figure 4-1: Areas covered by the two (coarse grid and fine grid) models. BC7 BC2 BC6 BC5 BC3 / BC1 I BCg8 siou Figure 4-2: Cedar/Ortega/St. Johns River grid with open boundary locations. Figure 4-3: Cedar/Ortega/St. Johns River bathymetry. Depths are in meters. Hourly water level and salinity data from the St. Johns River (supplied by the St. Johns River Water Management District) were used to define the boundary conditions at the north and south open boundaries of the St. Johns River (BC7, BC8). As an example, Figure 4-4 shows the time-series and the cumulative distribution of the measured water surface elevation data. The mean tide range is approximately 0.4 m. 0.6 0.4 E- 0.2 0 a3 0 -0.2 -0.4 -0.6 -0.8 '7II 90 100 110 120 130 140 150 160 Time, days 100 / I 80 60-/ a) 140 20 0 r I I I I 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Water elevation, m Figure 4-4: Cedar/Ortega/St. Johns River open boundary condition (BC7) during 2001 showing water surface elevation time-series and cumulative distribution. Small creeks (Williamson Creek, Butcher Pen Creek, Big Fishweir Creek and Fishing Creek denoted on Figure 4-2 as BC6, BC5, BC3 and BC4, respectively) minimally affect the flow in the larger estuary. Hence, instead of using "open" boundary conditions there, they were defined in terms of sink/source cells for specifying the flow and sediment flux conditions at the heads of these creeks. The relevant boundaries BC2BC6 are shown in Figure 4-5. The boundary condition time-series were supplied by the St. Johns River Water Management District. E 50 40 30 20 10 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 Discharge, m3/s Figure 4-5: Cedar/Ortega/St. Johns River tributaries discharge, cumulative distribution. Measured salinity time-series at open boundaries BC7 and BC8 were used. The model was run for a thee-week period to establish the salinity field (model "spin-up"), defined by these time-series. The bottom roughness z0 was chosen as 0.04 m throughout. The kinematic viscosity and molecular diffusivity were set to 10-6 and 10 l/s, respectively [equations (2-1) (2-4); Section 2.2]. The period of simulation corresponded to April 25, 2001 through May 30, 2001. A time-step of 3 s was used. The purpose of the simulation run was to generate the flow, salinity and suspended sediment time series at the downstream boundary of the Cedar River model (near the Cedar/Ortega confluence), and also to establish a (conservative) salinity field over the estuary. These outputs were then used to generate the initial and boundary conditions for the Cedar River model. 43 4.1.2 Cedar River Model Setup The Cedar River grid was also horizontally Cartesian and vertically ca-stretched. Its horizontal dimensions were 160 by 450 cells, each cell representing an area of 15 mn by 15 mn, and also used six horizontal layers. The grid and bathymetry are shown in Figure 4-6 and Figure 4-7, respectively. EFDC Cedar River grid Trap 1 (CP1) Williamscn Creek Trap 3 (CP.) Trap 2 (CP2) Fishing creek Figure 4-6: Cedar River model grid. Butcher Pen Creek Cht rifrt hytly 06 07 08 09 1 1.1 1. Cedar River bathymetry. Depths are in meters. 02 03 04 05 Figure 4-7: The boundary conditions at the upstream ends of the Butcher Pen Creek, Fishing Creek, Williamson Creek and the Cedar River were forced by establishing sink/source cells with the given discharge time-series (Figure 4-5). The downstream open boundary was represented by the water surface elevation forcing time series (Figure 4-8) generated by the Cedar/Ortega/St. Johns River model. The same physical boundary also served as a boundary condition for sediment concentration, which was defined based on the water sample data (May 17, 2001). The initial salinity field was generated by approximating the salinity field from the Cedar/Ortega/St. Johns model, which helped in decreasing the time needed for model "spin-up" required to establish a conservative salinity field. The bottom roughness coefficient, viscosity and diffusivity were unchanged from the values used in the Cedar/OrtegaiSt. Johns River model. Figure 4-8 shows a reasonable agreement between the measurement and simulation of tide. Note that the "measured" time-series was derived by averaging the tides at stations TG2 and TG3 by taking the time lags into consideration, in order to represent tide at the open boundary, which occurred in-between the two tide stations. The mean range was 0.52 im for the measured tide and 0.50 mn for the calculated one. Figure 49 shows the predicted water surface elevation plotted over a short period of time at three control stations (cells): 1) at the confluence of Butcher Pen Creek and the Cedar River), 2) at the upstream-most cell of the Cedar River grid, and 3) at the cross-section in Cedar River at its confluence with the Ortega. The time lag between the upstream control station and the confluence station is equal to approximately 7.5 min, which is consistent with the registered time lag in the measured water level data, considering the distance (5.4 kin) between the stations. The simulated discharge data were compared to discharge obtained by the ADCP. The measured and simulated discharges in the Cedar River at the confluence are plotted on Figure 4-10. Figures 4-11 and 4-12 show similar results for the north and south crosssections of the Ortega River, respectively. The latter two simulations were derived from the Cedar/Ortega/St. Johns River model. In general, the simulated discharge appears to be in a reasonable agreement with measurement, especially considering measurement errors (see Appendix A). 0.5 . measured (avg of 2 stations) calculated 0.4 0.3 0.2 0.1 oI E 0 F -0.1 -0.2 -0.3 -0.4 -0.5 i i i i i 122 124 126 128 130 132 134 Figure 4-8: Measured and simulated water level variations at the downstream boundary of the Cedar River. 48 0.15 1 - Butcher Pen creek upstream confluence 0.1 E - 0.05 0 -0.05 122.3 122.305 122.31 122.315 122.32 122.325 122.33 122.335 122.34 122.345 122.35 Time, days Figure 4-9: Water surface elevation at three control points in Cedar River. 100 x Simulated discharge x ADCP measured discharge x ADCP estimated total discharge 50 - 50 -100 _150J 137 137.1 137.2 137.3 137.4 137.5 137.6 137.7 137.8 137.9 138 Time, days Figure 4-10: Measured and simulated discharges through the Cedar River cross-section. 0 20 -40 -60 137 137.1 137.2 137.3 137.4 137.5 137.6 137J Time, days Figure 4-11: Measured and simulated discharges through the (north cross-section of the confluence). 80 Simulz ADCP 60 x ADCP 40 20 E 0 -20 , 05 i Ortega River cross-section 137.4 137.5 Time, days Figure 4-12: Measured and simulated discharges through the Ortega River cross-section (south cross-section of the confluence). 4.2 Sediment Transport Model Setup and Calibration 4.2.1 Sediment Transport Model Setup For running the sediment transport model the initial suspension concentration was set to 5 mg/i, the average value for the sediment concentration in the Cedar/Ortega/St. Johns River estuary. For the upstream boundary conditions in Cedar River, Williamson Creek and Butcher Pen Creek supplied by the St. Johns River Water Management District were used. It should be pointed out that at the downstream boundary of the Cedar River, the outputted values from the Cedar/Ortega/St. Johns River were significantly lower (-5-12 g/l), than the values, ranging between 8 and 57 mg/i (with an additional, exceptional value of 101 mg/1 in one case), obtained from water sampling on May 17, 2001. Depth-averaged TSS concentration series simulated by the coarse grid model at the Cedar River cross-section near the confluence is shown in Figure 4-13 and commensurate collected samples at the same location are shown in Figure 4-14. Sediment concentration at the Cedar River cross-section near the confluence 137 137.1 137.2 137.3 137.4 137.5 137.6 137.7 137.8 137.9 138 Time, days Figure 4-13: Depth averaged TSS concentrations at the Cedar River cross-section near the Cedar/Ortega confluence simulated by the coarse grid model (May 17, 200 1). 137.5 137.55 137.6 137.65 137.7 137.75 Time, days 137.8 137.85 137.9 Figure 4-14: Depth averaged TSS concentrations from the water sample data, collected at the Cedar River cross-section near the Cedar/Ortega confluence (May 17, 200 1). * 137.95 The above discrepancy between the simulated and measured concentrations was found to be due to the low concentrations predicted at the head boundaries of the Cedar River and the creeks. These boundary conditions, supplied by the St. Johns River Water Management District, were not verified. The problem was unfortunately realized towards the end of the present study. It was however felt that rerunning the sediment transport calculations for the Cedar River was not necessary, because the trapping efficiency results, described later, rely on relative rather than absolute values of the sediment flux. Thus the conclusions of the study were not affected. Since variation in the sediment concentration with time was small (8-9 mg/I), compared to the increase in concentration with depth (as found from the water sampling analysis), the sampled data were averaged, and a representative vertical profile of concentration with linearly distributed values from 14 g/l in the top layer to 27 g/l in the bottom layer, was used to set the open boundary condition at the in the Cedar River. 4.2.2 Bed Erosion The bed erosion function (lines representing erosion rate as a function of the bed shear stress) required for sediment model code is shown in Figure 4-15. It was based on laboratory experiments by Gowland (2002) using mud samples collected from the Cedar and the Ortega Rivers. This function was used for both models, i.e., coarse and fine grid. X 10 I III Measured data - Linear fit 0.8-7 Z* z0.6 U) 0 u 0.4 0.2 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 Bed shear stress, Pa Figure 4-15: Bed erosion rate function obtained from laboratory experiments on mud from the Cedar/Ortega Rivers (after Gowland, 2002) 4.2.2 Settling Velocity and Deposition The settling velocity model (Section 2.4) was calibrated using the data obtained from laboratory settling column tests using sediment from the site (Gowland, 2002). Some model tests were also carried out it against data available in the literature, the results of which are given in Appendix B. The model did not function for the values of the dissipation parameter G on the order of magnitude of 10.3 Hz and less, because very low turbulence levels caused the particle to grow infinitely large. This was due to the model formulation, in which particle size is dependent on a level of turbulence. It should be noted that in reality the volumetric concentration 0 c D cannot exceed unity, by definition, hence the constraint p, DI, D I, To calibrate the model against the laboratory data (Gowland, 2002) the value of G 10-2 Hz was used, in order to satisfy the above constraint and, at the same time, to simulate a near-quiescent situation (the settling column being a quiescent environment). The estimated range of G for Cedar River estuary was found to be within 0.5 to 10 Hz [based on equation (2-30), which gives a relation between the flow velocity, shear stress and dissipation parameter]. The concentration in the Cedar River (from the water samples, collected in May 17, 200 1) was found to be within 8 to 5 7 rng/l; however for the modeling purposes the range of concentrations from 0.1 to 100 was selected for convenience. Figure 4-16 thus obtained qualitatively resembles the Dyer diagram (Figure 4-16). The dissipation parameter was related to the shear stress using equation (23 0) considering the mean flow of 10 cm/s. E 2oo0150 o_ 10050, O> 0 0.4 -. o3104 0.6 0.8 102 Shear stress N/m2 1 101 Concentration mg/I Figure 4-16: Calculated floc size as a function of shear stress and concentration. Figure 4-17 shows settling velocity as a function of concentration, based on the laboratory settling column data using sediment from the site (Gowland, 2002). The curve is described by the equation W ( C (4.1) with the parameters a, h, n and n? set to 0.035, 2.0, 3.5 and 2.75, respectively. 10 . 0Measured data ................. Meaure .data ............... ; . .... .... ... F itte d c u rv e 1 0 3 . . . .. 10 ' -10-5 E 0 ..........~~~ ~~~~ . .. .. .. . . .. . ..... .. 10 10-1 100 101 102 Concentration, kg/m3 Figure 4-17: Settling velocity curve based on laboratory tests in a settling column using sediment from the Cedar River and vicinity (after Gowland, 2002). For calculation purposes, the value of the fractal diameter, nf [equation (2-26)], was taken as 2.3. By fitting the settling velocity predicted by the model to the curve given by Figure 4-17, the parameters (k,,, kh, p and q) for the settling calculation velocity were found. The exponents p and q were found to be 0.7 and 0.5, respectively, and the growth/breakup efficiency coefficients k,, = 10.3 and k, = 16.8 .103. The values of these coefficients are of the same order of magnitude as those of Winterwerp (1998) (k, = 14.7 and k, = 14.0. 10' ). Fluid properties were selected as p, = 1,020 kg/m3 and v = 10-6 m2/s. 4.3 Trapping Efficiency Analysis 4.3.1 Treatment Plan As described in Section 1 .2, the effects of two sediment treatment sites were to be tested. For the present purpose the locations of the sites (I and 2) were changed (Figure 4-6). Each site was tested with four assumed trapping efficiencies: 0% (no trapping), 30%, 60% and 90%. The maximum efficiency (90%) is in part based on the estimated 85% for TSS removal by Wet Detention Systems (WDS) in Florida; see Table 4- 1. Table 4-1: TSS removal efficiencies of treatment systems in Florida (after Harper, 1997) Treatment system Estimated TSS removal efficiency (c Dry Retention 60-98 Off-Line Retention/Detention 90 Wet Retention 85 Wet Detention 85 Wet Detention with Filtration 98 Dry Detention 70 Dry Detention with Filtration 60-70 Alumn Treatment 90 For that purpose, the resulting (calculated) settling flux (total mass of sediment passing the cross-section of the estuary in a unit of time) values at the Cedar River open boundary were compared to determine the potential effect of trapping sediment near the upstream end of the Cedar River (Site 3) on deposition downstream, where contaminated sediments derived from upstream- Cedar River tend to deposit. It should be noted that in the Cedar River the direction of the water flow changes with flood and ebb tides; hence the ebb tide is the only time when there is a sediment flux out of the river. 4.3.2 Sediment Trap Setup The sediment trap at the treatment site typically is a water detention (i.e., temporary retention) pond. By diverting river flow into the pond where flow velocities are small, a major portion of suspended sediments will typically deposit. Such systems can also be effective for storm water treatment when the bulk of the solids is carried with the first flush, as they can be intercepted and given a sufficient residence time to allow them to deposit. While some treatment facilities may require drainage pumps, others are strictly gravity flow systems. If the water is high in nutrients, the facility may include a vegetated wetland area that will absorb the nutrients in the water before it is discharged into the receiving waters. The concern for the Cedar River treatment system was to provide as much treatment as possible; hence the effectiveness of the facility was defined by the area available for it. Due to modeling limitations and related complications in representing the site as a water body with channelized flow diverted into it, site representation in the model was simplified. Accordingly, a function was implemented that decreased the sediment flux bypassing the grid cell by a pre-defined percentage. The channel cross-section, where the treatment site would be located, was represented by cells having such a sediment removal function (in terms of the percentage by which the effluent sediment load, leaving the site, is reduced with respect to influent load entering the site). 4.3.3 Effect of Trap Efficiency on Settling Flux Downstream Cedar River model runs were run without and with the sites in place (Figure 4-6), each for the selected four removal ratios (0%, 30%, 60%, and 90%). The model was run for three days, during May 16-18, 2001. Four output control points (CPI1-CP4) were selected (Figure 4-6). CP1I-CP3 corresponded to the sites and were placed just upstream of a site to measure sediment flux into the site, and CP4 was the control point just upstream of the open boundary, for monitoring trapping influence at the downstream end. Net sediment fluxes at the control points averaged over three semi-diurnal tidal cycles (the second cycle on May 17 and two cycles on May 18, 2001) are presented in Table 4-2. As seen in this table, it can be inferred that Sites I and 2 in the upstream portion of the Cedar River would have a small effect on sediment transport at the lower end of the Cedar River. In contrast, Site 3 can be considerably more effective. The reason for these differences appears to be that the majority of sediment load is derived from Williamson and Butcher Pen Creeks, rather than the Cedar River. Table 4-2: Comparison of sites with different removal efficiencies with a no-trapping scenario. Trap efficiencies ( )Net sediment flux g/s Resulting efficiency at tile Tr ITrap 2 Trap 3 C13I CP2 CP3 CP4 con~fluen~ce (o 0 0 0 6.7 8.9 11.4 14.1 0.0 30 0 0 6.7 7.9 11.0 13.8 2.1 60 0 0 6.7 7.7 9.9 13.6 3.5 90 0 0 6.7 7.0 9.6 13.1 7.1 0 30 0 6.7 8.9 10.1 12.3 12.8 0 60 0 6.7 8.8 8.2 10.9 22.7 0 90 0 6.7 8.7 7.6 9.2 34.7 0 0 30 6.7 8.9 11.4 10.6 24.8 0 0 60 6.7 8.8 11.4 7.9 44.0 0 0 90 16.7 8.8 11.2 15.8 58.9 30 30 30 6.7 7.8 7.2 6.9 51.1 60 60 60 6.7 6.0 4.1 2.9 79.4 90 90 90 6.7 2.8 1.1 0.3 97.8 The above observations are further highlighted in Table 4-3, by taking the cases of no entrapment and 30% entrapment (which may closer to a realizable efficiency), and seeing the effect in the confluence area. From the table it appears that: 1) Any treatment facility upstream of Williamson and Butcher Pen Creeks, as presently envisaged, will not be effective in reducing sediment loading in the confluence area, 2) treatment downstream of Butcher Pen will have measurable, but possibly not significant effect, and 3) more than one treatment site may have to be developed. In the event that a downstream treatment site cannot be constructed, dredging a trap in the river bed at that site should be 60 considered. Such an action should preferably be coupled with a one-time dredging of the confluence area to remove soft sediment deposit there. Table 4-3: Summary of the effect of treatment on sediment load in the confluence area Upstream Mid-stream Downstream Load reduction at the confluence (%) Cedar Cedar Cedar Treatment -2 Treatment -13 Treatment -25 Treatment Treatment Treatment -51 CHAPTER 5 CONCLUSIONS 5.1 Summa A study of the effectiveness of fine sediment trapping in the Cedar River estuary in north Florida was carried out. A combined three-dimensional hydrodynamic and sediment iriodel was set up and calibrated for this estuary. The original model (EFDC) was improved for fine sediment settling velocity calculations, by accounting for floe growth and break-up processes due to turbulence. The effect of selected sediment treatment sites or traps with different efficiencies (and placed in different locations upstream) on sediment transport downstream was examined. 5.2 Conclusions The following are the main conclusions of this study: 1. Simulated discharge and tidal variations in the Cedar River were found to agree reasonably well with measurements. 2. The settling velocity calculation routine was found to be applicable to conditions when the flow is turbulent, but not in near-quiescent waters, i.e., when the energy dissipation parameter has low values. 3. It appears that fine sediment trapping in the upstream reach of the Cedar River would have only a minor effect on sediment transport downstream near the confluence of the Cedar and Ortega Rivers. This is so because a major part of suspended sediment flux downstream appears to arrive there from creeks (especially Butcher Pen and Williamson) that flow into the middle reach of Cedar River. 4. Sediment entrapment closer to the confluence of the Cedar and Ortega rivers appears to be able to measurably reduce sediment transport to the confluence and, therefore, can be expected to lower the flux of contaminants out of Cedar River. 5.3 Recommendations for Further Work Further development of the settling velocity model is required, in order to extend the calculation to settling in near-quiescent water. Traps simulation should be made more realistic by incorporating the mechanics of an actual retention/detention pond in the model. APPENDIX A WATER DISCHARGE ESTIMATION BASED ON ADCP Water discharge must be calculated for each ADCP transect. Due to the inability of the ADCP used to record measurements close to the water surface and the bottom, as well as in the shallow near-bank areas, a method must be used to account for the loss of coherent signals for these blank zones in an approximate way. The following uses the method suggested and used in RD Instruments, WinRiver (software designed by RD Instruments for analysis and visualization of the ADCP data) and is described in a help system provided with the software The required total discharge (Q,,) from the instrument consists of measured ( ,.. ) and estimated ( Q,, ) values: Q-a = Qesl + Q, (A. 1) The estimated discharge, which must be added to the measured value, consists of four components: top (layer close to the surface), bottom (layer close to the bottom), right and left (discharge in the zones close to the bank, where ADCP data are usually not taken because of shallow water). The "estimated" discharge is then calculated as: Q -S, = Q,,p + Qho,,,,, + Q,,, + Q,, (A. 2) For calculating the top and bottom discharges two (user-definable) methods can be used: Constant and Power. The Constant Method assumes that the velocity is constant in the top/bottom layer and is equal to topmost/bottommost successfully measured acoustic bin. The Power Method assumes a power-law velocity (u) profile in the vertical z ) direction: it = r. z (A. 3) in which the exponent p is user-defined with a default value of 1/6, and the proportionality constant r is found by fitting the power-law profile to the measured points. For calculating the right/left discharge the following formula is used: Q = c.D.H u,,,' (A. 4) where c is a user-defined coefficient equal to0.35 for a triangular near bank bottom shape (default) and 0.91 for a rectangular near bank bottom shape; D is a distance to bank (defined in a data collection process and obtained from the ADCP data files); H is a water depth of the leftmost/rightmost measured ensemble; and u,, is velocity averaged over the user-defined number of leftmost/rightmost ensembles. APPENDIX B SETTLING VELOCITY AND FLOC SIZE CALCULATONS B. 1 Introduction In order to demonstrate the application of the settling velocity model described in Section 2-4 and to test the model against the data available in the literature, the following calculation tests were performed. B.2 Settling Velocity Calculations Wolanski et al. (1992) presented data on the settling of sediment from Townsville Harbor, Australia. A Plexiglas cylinder of 10 cm internal diameter and 140 cm height was used as a settling column. Turbulence could be generated in this column by oscillating 1 cm wide rings along the walls, spaced 2 cm apart. Two sets of data were obtained: in quiescent water, and with rings oscillating. Quiescent water can be characterized by very low values of dissipation parameter G. First, model predicted settling velocity was fitted to the data in oscillating flow based on equations (2-29), (2-36) and floc aggregation coefficient in a form similar to CH equation (4- 1), i.e., k,=" (b +) The parameters k, kh, p and q were determined in this way; p and q were found to be 0.6 and 0.45, respectively, and. k, = 8.7 and k,=19.1.10' for n=0.87 and b=1.96. A representative value of the dissipation parameter was found to be G 1.3 Hz. Then G 10 2 Hz was used to represent quiescent water and the corresponding velocity curve was plotted (Figure B-1). In Figure B-i, the simulated curve based on measurement in oscillating water indicates a reasonably good match with data points. However, measurements in quiescent water are not predicted as well. This is believed to be due to the fact that, as noted in Section 4.2.2, the model does not perform well for low values of dissipation parameter G (i.e., in the absence of turbulence). * *t( Lab results (quiescent water): : .... .. ... ... x % Lab results (Max. velocity=0.09 m/s) 10- 31 I.I.. I.. .L L ...L ; ; ; ; r ; 10-1 100 101 102 Concentration g/Il Figure B-i: Settling velocity calculation test results, and comparison with data of Wolanski et al. (1992) using sediment from Townsville Harbor, Australia. 102 B.3 Particle Size Calculations In steady flows and with given sediment properties, flocs tend to have a narrow size distribution and may be assumed to have an equilibrium size defined in terms of, for example, the median diameter. The equilibrium size condition implies that the growth and breakup processes balance each other. Thus, flocs that are smaller than the equilibrium size would have growth dominating over breakup, and for larger flocs the breakup process would be dominant. As a result the floc size tends to fluctuate around its equilibrium value. The model was tested against the floc size data published by Winterwerp (1998) from two settling column tests using sediment from the Ems-Dollard River area in The Netherlands. Particle sizes were measured using a Malvern particle sizer. For simulation purposes the parameter values in Table B- 1, plus those provided by Winterwerp, were used. The initial particle size was taken as 4 jim, as measured by Winterwerp. Coefficients k, = 14.7 and k,, = 14.0 103 were selected. Comparisons between simulations and data are shown on Figure B-2. The values of concentration and dissipation parameter used are given in Table B-1. The resulting curves, which lead to equilibrium sizes, appear to be the same as those of Winterwerp (1998). Floc size is seen to grow with time until it reaches an equilibrium value (there is an equilibrium particle size for given concentration and dissipation parameter) and remains the same beyond that point. E / / 10, 10' id id. 16 i. Time, s Figure B-2: Floc growth with time measured and predicted for River Ems-Dollard mud (Winterwerp, 1998). Table B-1: Data from settling column tests with Ems-Dollard mud. c G Test No (kg/m3) (Hz) T-73 1.21 81.7 T-69 1.17 28.9 REFERENCES Dyer, K. R., 1989. Sediment processes in estuaries: Future research requirement. Journal of Geophysical Research, 94(C 10), 9489-9498. Ganju, N. K. 2001. Trapping organic-rich sediment in an estuary. M.S. Thesis, University of Florida, Gainesville. Gowland, J. E., 2002. Laboratory experiments on the erosional and settling properties of sediment from the Cedar/Ortega River system. Report UFL/COEL-CRJ2002/001, Coastal and Oceanographic Engineering Program, Department of Civil and Coastal Engineering, University of Florida, Gainesville. Hamrick, J. M., 1992. A three dimensional environmental fluid dynamics computer code: Theoretical and computational aspects. Special Report No 317, Applied Marine Science and Ocean Engineering, Virginia Institute of Marine Science, Gloucester Point, VA. Hamrick, J. M., 1996. User's manual for environmental fluid dynamics computer code. Special Report Special Report No 331, Applied Marine Science and Ocean Engineering, Virginia Institute of Marine Science, Gloucester Point, VA. Harper, H. H., 1997. Pollutant removal efficiencies for typical stormwater management systems in Florida. Proceedings of the Biennial Stormwater Research Conference, Southwest Florida Water Management District, Tampa, FL, 6-19. Huang, H., 1994. Fractal properties of floes formed by fluid shear and differential settling, Physics of Fluids, 6(10), 3229-3234. Levich, V. G., 1962, Physicochemical hydrodynamics, Prentice Hall, Inc. McAnally, W. H., Mehta, A. J., 2000. Aggregation rate of fine sediment. Journal of Hydraulic Engineering, 126(12), 883-892. McAnally, W. H, and Mehta, A. J., 2001. Collisional aggregation of fine estuarine sediment. In: Coastal and Estuarine Fine Sediment Processes, W. H. McAnally and A. J. Mehta eds., Elsevier, Amsterdam, 19-40. Mellor, G. L., and Yamada, T., 1982. Development of a turbulence closure model for geophysical fluid problems. Reviews in Geophysics and Space Physics, 20, 851875. National Research Council, 2001. A risk-management strategy for PCB-contaminated sediments. National Academy Press, Washington, DC. Parshukov, L. N., 2001. Effect of turbulence on the deposition of cohesive flocs. M.S. Thesis, University of Florida, Gainesville. RD Instruments, 1994. Transect: User's Manual for Broadband Acoustic Doppler Current Profilers, San Diego, CA.. Smolarkiewicz, P. K., 1983. A simple positive definite advection scheme with small implicit diffusion. Monthly Weather Review, 111, 479-486. Stoddard, D. M., 2001. Evaluation of trap efficiency in an estuarine environment. Report MRP-2001/003, Department of Civil and Coastal Engineering, University of Florida, Gainesville. Stolzenbach, K. D., Elimelich, M., 1994. The effect of density on collisions between sinking particles: implications for particle aggregation in the ocean. Journal of Deep Sea Research, 41(3), 469-483. Teeter, A. M., 2001. Clay-silt sediment modeling using multiple grain classes: Part I: Settling and deposition. In: Coastal and Estuarine Fine Sediment Processes, W. H. McAnally and A. J. Mehta eds., Elsevier, Amsterdam, 157-170. Van Leussen, W., 1994. Estuarine macroflocs and their role in fine-grained sediment transport. PhD dissertation, University of Utrecht, The Netherlands. Winterwerp, J. C., 1998. A simple model for turbulence induced flocculation of cohesive sediment. Journal of Hydraulic Research, 36(3), 309-326. Wolanski, E., Gibbs, R., Ridd, P., Mehta A., 1992. Settling of ocean-dumped dredged material, Townsville, Australia. Estuarine, Coastal and Shelf Science, 35, 473489. BIOGRAPHICAL SKETCH Vladimir Paramygin was born the first son of Tatyana and Alexander Paramygin in 1979 in Barnaul, Russia. In 1996 he graduated from Darby High School in Darby, Montana, and High School 69 in Barnaul, and entered the Department of Mathematics at the Altai State University in Barnaul. During his studies there he found his interests in various areas of applied mathematics and computer science. In spring of 2000 he received his bachelor's degree in applied mathematics and was admitted to the Graduate School of the University of Florida, to continue his academic work in the Coastal and Oceanographic Engineering Program of the Department of Civil and Coastal Engineering. |