UFL/COEL2002/002
EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN
ESTUARY
by
Vladimir A. Paramygin
THESIS
2002
LJNWERSITY OF
F L OR ID
Coastal & Oceanographic Engineering Program
Department of Civil & Coastal Engineering
433 Weil Hall P.O. Box 116590 Gainesville, Florida 326116590
pptpt~faf$i~re~i~f
L.
UFL/COEL2002/002
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, chairman 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
page
ACKNOWLEDGMENT............................................................................................... ii
L IS T O F T A B L E S .................................................................................... ..................... v
L IST O F F IG U R E S ........................................................................................................... vi
EFFECTIVENESS OF FINE SEDIMENT TRAPPING IN AN ESTUARY.................. viii
1 IN T R O D U C T IO N .............................................................................. ...................
1.1 P rob lem Statem ent .......................................................................... ................... 1
1.2 C ed ar R iv er ............................................... ......................................................... 2
1.3 O objective and T asks ....................................................................... .................... 3
1.4 O outline of C chapters ................................................................................................. 3
2 M ETH O D O F A N A LY SIS .................................................................................. ........ 5
2 .1 In tro du action ............................................ ........................... ............................... 5
2 .2 F low F field ........................................................ ....................... ..... .... 5
2.3 Sedim ent Transport ................................................................. .................... 10
2.4 Settling V elocity C alculation ........................................................... ..... ........... 11
2.4.1 B 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 CED A R RIV ER ESTU A RY ........................................................................... 20
3.1 Description of the Estuary................................................................................. 20
3.2 Tide, Waves, Current, Salinity and Wind Data..................... ......................... 22
3.2.1 T ide data ........................................................... .......................................... 22
3.2.2 Current data ........ ....................................23
3.2.3 W ind data ... ...................................................... ......................................... 24
3 .2 .4 W av e data ..................................................................................................... 2 6
3.2.5 Salinity data........................................ .............................. ........................... 27
3.3 D ischarge D ata ..................................................................................................... 28
3.4 Sedim ent C concentration ........................................................................................ 29
3.5 Bed Sediment Distributions ......................... ................. 30
iii
4 ASSESSMENT OF SEDIMENT TRAPPING EFFICIENCY ............................... 37
4.1 Flow Model Setup, Calibration and Validation .............................................. ...37
4.1.1 Cedar/Ortega/St. Johns Rivers Model Setup.......................... ........ .. 37
4.1.2 Cedar River Model Setup.............................................. .....................43
4.2 Sediment Transport Model Setup and Calibration...................... .......... .. 50
4.2.1 Sediment Transport Model Setup.............................. ..................... 50
4 .2 .2 B ed E erosion .................................................................. ........................... 52
4.2.2 Settling Velocity and Deposition ................................... ....................... 53
4.3 Trapping Efficiency Analysis.................................................... 57
4.3.1 Treatm ent Plan .......................................... .......................... 57
4.3.2 Sediment Trap Setup .................................................. 57
4.3.3 Effect of Trap Efficiency on Settling Flux Downstream ............................. 58
5 C O N C LU SIO N S ........................................................... .................................. 61
5 .1 S u m m ary .............................................................................. ................. .......... 6 1
5.2 C onclu sion s ............................................................. ............. . ...... ............... 6 1
5.3 Recommendations for Further Work.............. ............................................. 62
APPENDICES
A WATER DISCHARGE ESTIMATION BASED ON ADCP................................... 63
B SETTLING VELOCITY AND FLOC SIZE CALCULATIONS ............................65
B Introduction .......................................................................... .................... 65
B .2 Settling V elocity Calculations ............................................ ...................... 65
B .3 Particle Size Calculations .................................................... ... .................. 67
R EFE R EN C E S ........................... ...................................................................................... 69
BIOGRAPH ICA L SKETCH ........................................................... ..................... 71
iv
LIST OF TABLES
Table Page
31: Wind speed/direction distribution. Dominant speed/directions are highlighted... 25
32: Cedar River crosssection discharges, May 17, 2001 .......................................29
33: Sediment concentrations from water samples, May 17, 2001........................... 30
34: Statistical values associated with bed sediment distribution............................... 32
41: TSS removal efficiencies of treatment systems in Florida (after Harper, 1997)... 57
42: Comparison of sites with different removal efficiencies with a notrapping
scene ario ........................................................................... ..... ................ ......... 5 9
43: Summary of the effect of treatment on sediment load in the confluence area...... 60
Bl: Data from settling column tests with EmsDollard mud..................................... 68
LIST OF FIGURES
Figure Page
21: Effect of sediment concentration and fluid shear stress on the median floc
diam eter (D yer, 1989). .......................................................... ..................... 13
22: Relationship between settling velocity and floc diameter in still water, based on
equation (2.29). .................................................................... ..................... 17
31: Cedar/Ortega Rivers estuary, (within the light rectangular area), aerial photo,
M ay 19 9 8 .......................................................... ..................... ............ ..... ..... 2 0
32: Cedar/Ortega Rivers data collection and sediment treatment (Wet Detention
System ) sites 1 and 2 .................................................. ............................... 2 1
33: Tidal ranges at stations TG TG3. Cumulative frequency distribution based on
record obtained during 09/29/0010/18/01............................ .................... ... 23
34: 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 .................... ... ...... .............. ...................................... ................ .. 24
35: Wave spectrum based on 10/02/01 25/04/01 record at WGC............................. 26
36: Significant wave height based on spectral analysis. .......................................... 27
37: Cumulative distribution of the significant wave height at the mouth of the Ortega
River in the St. Johns River.................... ........................................................... .. 27
38: Cumulative distribution of salinity at stations TG TG3; 10/27/0011/26/00...... 28
39: Cedar River crosssection discharge and Cedar/Ortega tide data, May 17, 2001. 29
310: Moisture content distribution (%). Based on 1998 sampling................................ 33
311: Organic content distribution (%). Based on 1998 sampling. ............................ 34
312: Solids content distribution (%). Based on 1998 sampling................................ 35
313: Thickness of soft deposit in the study area based on core thicknesses in 1998
sam pling. ............................................................................... 36
41: Areas covered by the two (coarse grid and fine grid) models........................... 38
42: Cedar/Ortega/St. Johns River grid with open boundary locations...................... 39
43: Cedar/Ortega/St. Johns River bathymetry. Depths are in meters........................ 40
44: Cedar/Ortega/St. Johns River open boundary condition (BC7) during 2001
showing water surface elevation timeseries and cumulative distribution.......... 41
45: Cedar/Ortega/St. Johns River tributaries discharge, cumulative distribution....... 42
46: Cedar River m odel grid. ........................................................ .................... 44
47: Cedar River bathymetry. Depths are in meters. ............................ ............. 45
48: Measured and simulated water level variations at the downstream boundary of the
C edar R iver. ........................................ ........................................................ 47
49: Water surface elevation at three control points in Cedar River ........................ 48
410: Measured and simulated discharges through the Cedar River crosssection........ 48
411: Measured and simulated discharges through the Ortega River crosssection (north
crosssection of the confluence)......................... ................ ....................... 49
412: Measured and simulated discharges through the Ortega River crosssection (south
crosssection of the confluence) .................................................. ....................... 49
413: Depth averaged TSS concentrations at the Cedar River crosssection near the
Cedar/Ortega confluence simulated by the coarse grid model (May 17, 2001).... 51
414: Depth averaged TSS concentrations from the water sample data, collected at the
Cedar River crosssection near the Cedar/Ortega confluence (May 17, 2001)..... 51
415: Bed erosion rate function obtained from laboratory experiments on mud from the
Cedar/O rtega Rivers ................................... ....................... ........................ 53
416: Calculated floc size as a function of shear stress and concentration................... 55
417: Settling velocity curve based on laboratory tests in a settling column using
sedim ent from the Cedar River and vicinity. ..................................................... 56
B1: Settling velocity calculation test results, and comparison with data of Wolanski et
al. (1992) using sediment from Townsville Harbor, Australia ........................... 66
B2: Floc growth with time measured and predicted for River EmsDollard 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 submodel 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 sedimentbound 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 organicrich sediment can increase contaminant loads in these waters,
because contaminants such as polychlorinated biphenyls (PCBs) and polyaromatic
hydrocarbons (PAHs) are preferentially bound to organic (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 online or offline. One
example of an online trap is a dredged trench along the submerged bottom, which
reduces the flow velocity and causes the material to settle there. An offline 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, 2001). 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 sedimentbound 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 km east of the Cedar
River near its headwaters, adjacent to municipal storm drains and drainage ditches.
There, fire destroyed several tanks storing high concentrations (4,425 ppm) of PCBladen
oils and other materials. It is believed that a combination of damage to the storage tanks
and the firefighting 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. Finegrained sediments play an
important role in these processes. Due to relatively strong currents, fine sediments, which
are mixtures of clay and siltsized 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 offline 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 PCBladen, organicrich 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:
1) Use of data to characterize the nature of flow. Data included tidal elevations,
current velocities, wind speed and direction, salinity, streamflows 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 threedimensional,
hydrostatic flow model, as well as a compatible sediment model.
2.2 Flow Field
The coordinate system of the model 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 (ogrid). 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
D,(m,m,,Hu)+ ,. (m,Huu)+ a ,(myHvu)+ a. {mm.iwu) f.mm, Hv
=m,Hm (p +p ,,,,, + )+ m,,(xz;. +z, H)p+ l mm,, u) (2.1)
H
+. HA x x HAa uj mm,cD,(u +v2 )12
mx ) \ m, y
(mm ,,Hv)+ 9,. m,.Huv)+ 0 (mrHw)+ (mxm,,wv) fmm Hu
=m xH,.I(p+p,,,,, +p )+ m, (.,z', + z.,,H).p+. nim m,Am .v (2.2)
+a., HAH.d v +a." HAHaYV _nm) cPD P(u2 +v2)12 V
m,m,,f, = mxm,,f Ul,mx + vaxm, (2.3)
)x:>,)= A,.H 'a(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; m, and nm, 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; 0 = 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); A, 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 o )po' (2.5)
where
p = p(p,S,T) (2.6)
and p and p, are the actual and reference water densities, respectively, and b is
buoyancy.
The threedimensional continuity equation in the model is
a, (rnm, H)+ a,(m,,Hu)+ ,,(m_,Hv)+ a (mwm,,w)= QH (2.7)
and the corresponding verticallyintegrated form of the continuity equation is
a,(mx mH)+ a(m,,/u)+a y (mx v)= QU (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)+ (mHuS)+ ,(mHvS)+ (mwS)= a:(mH 'A,zS)+Q, (2.9)
a,(mHT)+ a,(mHuT)+3a,(mHvT)+ ,(mwT)= (mH 'A,,aT)+ Q, (2.10)
where Q, and Q, are source and sink terms, respectively, which include subgrid scale
horizontal diffusion and thermal sources and sinks, and A, is the vertical turbulent
diffusivity.
Two transport equations determine the turbulent intensity and turbulent length scale as
follows:
M, m mY.Hq2)+3, (mHuq2)+ ,, (mxHvq2)+ (mxm,wq2)
A, 2 Hq3
=3, mxmyI q 22mxm, B (2.11)
H B1l
+2mm ((dyu)2 +(. v)2)+qc7pD(lu 2 +v gK,.+:b)+Qq
3, (mm,.mHq2 1)+ 3 (m Huq2l)+ a, (mxHvq2l)+ 3 (m.m,,wq21)
= nmxm AdL (q2Hq1) mm r+E2( H 2 +E3 kH( 1) (2.12)
H B, kHz kH( (z)
+mxmyE, l ((3u)2 +(a3v)2)+gK,3b+7cpD,(u2+v2) +Q,
where (E,,E,,E3)= (1.8,1.33,0.25) are empirical constants, Qq and Q, represent
additional sourcesink terms, 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 rq,.
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, )= (r,,,)= c JU,, + 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 (0.8+0.065 U2 + V,) (2.14)
AOW
where P'i and P,, are the air and water densities. At the bed, the shear stress components
are considered to be related to the nearbed or bottom layer velocity components as
follows:
( . ,,; )= rh., r,,)= cC, U ` +v (u,, v,) (2.15)
where ul and v' are the bottom layer velocity components and the bottom stress
coefficient is
ch,= (2.16)
InA
which assumes that the nearbottom velocity profile is logarithmic. In equation (2.16) K
is the von Karnan constant, A, is the dimensionless thickness of the bottom layer, and
zo = is the dimensionless roughness height.
H
The vertical boundary conditions for the turbulent kinetic energy and length scale
equations are
2
q 2=BT, : z Z=1 (2.17)
2
q2 =B13 T : z =0 (2.18)
I=0 : z=0,1 (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
(mm,,HC)+ (m,,HuC)+ a, (mHvC) + (mm, ,wC) a, (m,m1,w,,C)
m m, mm K,, K (2.20)
=a, f H ,.K C +aY xHKa ",v +la mxm C+,c Y
mx. 'm ) 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
a, (mxmHS,)+ a, (m HvS,)+ a, (m.,HvS,)+ a (mm, mwS,)
aK(m rniwS= as + i, (2.21)
a (..( m ,w s ,)= a m, m a S + Q,'; + Qo;
where S, represents the concentration of the jth sediment class. Sourcesink are
represented by two terms: an external part, which would include point and nonpoint
source loads, and an internal part, which could include reactive decay of organic
sediment, or exchange of mass between sediment classes when floc growth and breakup
are simulated.
The vertical boundary conditions for equation (2.21) are
K
K K S wS=J :z=0
H o
H (2.22)
K
a.S wS = 0: z = 1
where Jo is the net water columnbed 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 finegrained
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 multiclass 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 21.
FLCC DIAM
200
SI \' CONCENTRATION
t10 0 104 \ ' .01,
SHEAI STRESS
dynes mn 2
Figure 21: 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 growth/breakup of fine sediment based on multiclass grain size
distribution, two and threebody 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 forms of those functions using
experimental data:
Collision diameter function:
F2 [ Ap (D,+D,,,) S D, ( T, j CEC (2.23)
Ap, ,,,' v So D, T, CECo
Collision efficiency function:
D, S 10.875 T CEC ____
,, n AD, so To CECo [,,,,,,,, T,
a = )e (2.24)
Ap, D, u, (D, + D,,, ) (2.24)
Ap,,,D,, ) v
where TI, is some function of the bracketed nondimensional terms; indices i and m
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;
To= 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; ,,,,, = shear stress
experienced by the m aggregate and CEC, CECo = actual and reference cation exchange
capacities, respectively.
Parshukov (2002) presents a multiclass 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 gridgenerated 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:
0 = P C = jfnD (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 f/ = 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 Apl
can be related to the fractal dimension of the particles using the formulation
Ap, = p, p ,, P,) (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 2.12.3 (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):
(1+aG)
Wb =W ( G) (2.27)
(l+bG )
where W, and W,o 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:
C I+a2GC
W= C tl+aG exp (a4 C 1 (2.28)
Cul 1+aG 3 C/
where C,, and C,, are massweighted average upper and lower reference concentrations,
respectively; and n, a, a2, 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 = (,P.")g3, D"' (2.29)
S18/8 p "P 1+0.15 Re0687 o
where a, ,f are the coefficient that depend on particle sphericity, these coefficients will
be taken as 1 (spherical particles) here; up =dynamic viscosity and Re = settling
p,,.W, D1,
Reynolds number,
Figure 22 shows the relationship between the floc diameter and the settling
velocity in still water, as described by equation (2.29).
D vs Ws
10,
102
:102' ..;*. : : .f 3 *y . ... .. : .: .
E .. nf=2.
.. . ... : . : ... . . ... . . n f= 1 : . .
100
10...2. . . ..... ... . . ...
101 102 103 104
D, mu
Figure 22: 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
du
G = ldz (2.30)
where e = energy dissipation rate of flow, v = kinematic viscosity of the fluid, 1t =
dynamic viscosity of the fluid, z = 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 eceGD3n2 (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 ere, CGD"'3D44/ = k cGD4''
dt 2 fnf p (2.32)
where
3e, ret D' D i3
k e =k, f (2.33)
2Th nfp,, nfpu ,
The rate decay of particle due to breakup is suggested as:
dD ae D G(2.34)
dD ae, DG DD uI 1 = k,G"+ (DDr)P D2q+ (2.34)
dt nf D, F
k, aeD DPj/D D,' (2.35)
"/ [F,) nf 1 F,
where p and q are the empirical parameters; e, 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:
D k c GD"' 3D4"n /k GI+ DD2q+1 (D Y (2.36)
dt nf p, 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 former 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. = 0102} and e,, =0.51.0 (Levich, 1962; O'Melia, 1985). Winterwerp (1998)
estimated k0 = 0{105 .
CHAPTER 3
CEDAR RIVER ESTUARY
3.1 Description of the Estuary
The Cedar River estuary is contained within Duval County in northeast Florida
Figure 31.
Figure 31: 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 RiverButcher Pen Creek and Williamson Creekalong with several smaller
tributaries. Another stream, Fishweir Creek also flow into the Ortega (Figure 32).
D
Site 1
Site 2
IQ
St Johns
River
WGC
o
O TG1
*0Se
oP
TG2
0
Ortega
River
ADCP
Transects
0 L>
/'
Ortega
River
1 kilometer
Figure 32: 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 m. Depths at the Cedar River vary from 0.3 m to 1.5 m 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 32 (TG1TG3). 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 33 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 33: Tidal ranges at stations TG1TG3. Cumulative frequency distribution based
on record obtained during 09/29/0010/18/01.
3.2.2 Current data
A tethered current meter (Endeco 174) was installed at location WGC (Figure 3
2). The data reported here is from an approximately 1 monthlong 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 34, the
current speed was below 30 cm/s 98% percent of the time, and below 25 cm/s 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, m/s
Figure 34: Current speed at station WGC (mouth of the Ortega in St. Johns River).
Cumulative frequency distribution based on record obtained during 02/05/0103/08/01.
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 31) provide information on the dominant wind
speeds and directions. These data indicate a potentially complex dependence of wind on
winddriven currents in the estuary, especially because portions of the waterway reaches
are lined by trees, while others have been cleared and developed.
Table 31: Wind speed/direction distribution. Dominant speed/directions are highlighted.
0 1
2 3 4 5 6 7 8 9 10 11 12 13 14 15 %
Speed,
m/s
24
48
72
96
120
S144
168
0
S 192
S216
240
264
288
312
336
360
%
occurrence
7s.i 11.5 I 20.8 I 22.7 18.2 I 9.6 6.6 3.2 0.0 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
4.8
23 148 298 377 321 181 145 97 27 10 4 1 1 4
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
11.5 20.8 22.7 1 18.2 9.6
6.6 1 3.2
0.0 1 0.0
1 5.1
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 35 is a spectral
representation of data. The variation of the significant wave height H,,, corresponding to
the modal period is shown in Figure 36.
The cumulative distribution of H,,,, shown in Figure 37, indicates that it did not
exceed 0.2 m. 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
L0.08
a 0.06,
0.04
0.02
160
140
120
0.4
100 0.3
80 0.2
60 0.1
Time, Days 40 Frequency, Hz
Figure 35: Wave spectrum based on 10/02/01 25/04/01 record at WGC.
0 15
015
0.05
40 50 60 70 80 90 100 110 120 130 140 150
Time, days
Figure 36: Significant wave height based on spectral analysis.
Wave data. HmO, from spectral analysis. Cumulative distribution.
100
90
80
70
60
50
Hm0, m
Ortega River in the St. Johns River.
40
30
20
10
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
HmO, m
Figure 37: 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 TG1TG3 during October27
November 26, 2000. Figure 38 shows the cumulative distribution of these data for each
station. The water is generally brackish; at TG1 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 TG3
I /
80
70 
60
E 50
SI I
40 
30
20
10 /
0 2 4 6 8 10 12 14
Salinity, ppt
Figure 38: Cumulative distribution of salinity at stations TG1TG3; 10/27/0011/26/00.
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
crosssections of the Cedar/Ortega River confluence area were selected for data
collection for almost a full semidiurnal 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 performance 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 32 presents the analyzed data for the
Cedar River crosssection (all transects were made at the crosssection of the Cedar River
near the confluence; see Figure 32). Positive discharge is directed west, and negative is
directed east.
Table 32: Cedar River crosssection discharges, May 17, 2001.
ADCP transect # Time Discharge, m3/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 crosssection was 1.0 m and maximum depth 1.4 m.
Figure 39 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 crosssection.
100 1 I I I
14 15 16 17 18
Time, hr
Water level at Main Bridge Cedar/Ortega River
19 20 21
Figure 39: Cedar River crosssection 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
10nnL
concentrations. Table 33 gives the data thus obtained. The concentrations are
characteristically low, and range between 8 and 101 mg/1.
Table 33: Sediment concentrations from water samples, May 17, 2001
e # Concentration, # Concentration, e Concentration,
Sample # Sample # Sample #
mg/l mg/1 mg/1
1 14 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, bedsampling data
supplied by the St. Johns River Water Management District were used to generate the
maps 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
determining 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:
V = , v, (31)
I =1
S= (32)
e
/1
in which v,= data point value, a = standard deviation of the dataset, and r,= 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 310), organic content
(Figure 311), solids content (Figure 312) and thickness of the deposit (Figure 313).
Measurement points are displayed as black dots and have values besides them. Also,
equalpercent contours are drawn to identify the areas where the percentage of
moisture/organics/solids is approximately the same.
Table 34 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 (3035%), whereas
the upstream reach of the Cedar River has a high solids percentage (2530%). 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 34: Statistical values associated with bed sediment distribution
Statc Moisture content Organic content Solids content
Statistic
(%) (%) (%)
Minimum 54 6 16
Maximum 84 51 46
Mean 76 21 24
Figure 310: Moisture content distribution (%). Based on 1998 sampling.
Figure 311: Organic content distribution (%). Based on 1998 sampling.
Figure 312: Solids content distribution (%). Based on 1998 sampling.
Figure 313: 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 41 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 m (Figure 42). The boundary conditions are
labeled BC1BC7. The grid was o stretched in the vertical direction with six horizontal
layers.
Figure 43 shows the bathymetry of the modeled domain. The Cedar/Ortega River
portion of the domain is typically shallow with 1.5 2 m depth in the channel thalweg,
and an average depth of m.
Area covered by
the Cedar River
Sgrid only
Area covered by the
Cedar/Ortega/St. Johns
grid only
Area covered by
both grids
Figure 41: Areas covered by the two (coarse grid and fine grid) models.
BC2
BC4
x
BC6
BC5
BC3
BCl
S
Figure 42: Cedar/Ortega/St. Johns River grid with open boundary locations.
BC7
gk^
Figure 43: 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 44 shows the timeseries and the cumulative distribution of the
measured water surface elevation data. The mean tide range is approximately 0.4 m.
0.6
0.4
0.2
0
0 0
0.2
M 0.4
0.6
0.8
90 100 110 120 130 140 150 160
Time, days
100
80
6 60
B 40
20
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Water elevation, m
Figure 44: Cedar/Ortega/St. Johns River open boundary condition (BC7) during 2001
showing water surface elevation timeseries and cumulative distribution.
Small creeks (Williamson Creek, Butcher Pen Creek, Big Fishweir Creek and
Fishing Creek denoted on Figure 42 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 BC2
BC6 are shown in Figure 45. The boundary condition timeseries were supplied by the
St. Johns River Water Management District.
E 50
40 /
30
20
10 ^
2 I
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 45: Cedar/Ortega/St. Johns River tributaries discharge, cumulative distribution.
Measured salinity timeseries at open boundaries BC7 and BC8 were used. The
model was run for a theeweek period to establish the salinity field (model "spinup"),
defined by these timeseries. The bottom roughness z0 was chosen as 0.04 m throughout.
The kinematic viscosity and molecular diffusivity were set to 106 and 108 T2/S,
respectively [equations (21) (24); Section 2.2]. The period of simulation corresponded
to April 25, 2001 through May 30, 2001. A timestep 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 ostretched.
Its horizontal dimensions were 160 by 450 cells, each cell representing an area of 15 m
by 15 m, and also used six horizontal layers. The grid and bathymetry are shown in
Figure 46 and Figure 47, respectively.
EFDC Cedar River grid
Tap 1 (CP1)
Hilliamson Creek
Trap 2 (CP2)
Butcher Pen Creek
Fishing creek
Figure 46: Cedar River model grid.
Mrap 3 (CP3)
02 03 04 05 06 07 08 Q9
1 1.1 1.;
Figure 47: Cedar River bathymetry. Depths are in meters.
I I I __
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 timeseries (Figure 45). The downstream open boundary
was represented by the water surface elevation forcing time series (Figure 48) 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 "spinup" required to establish a conservative salinity field. The bottom
roughness coefficient, viscosity and diffusivity were unchanged from the values used in
the Cedar/Ortega/St. Johns River model.
Figure 48 shows a reasonable agreement between the measurement and
simulation of tide. Note that the "measured" timeseries 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 inbetween the two tide stations. The
mean range was 0.52 m for the measured tide and 0.50 m for the calculated one. Figure 4
9 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 upstreammost cell of the Cedar River grid, and 3) at the crosssection 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 km) 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 410. Figures 411 and 412 show similar results for the north and south cross
sections 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
E 0
0.1
0.2
0.3 
0.4
0.5
122 124 126 128 130 132 134
Figure 48: Measured and simulated water level variations at the downstream boundary
of the Cedar River.
48
0.15
 Butcher Pen creek
upstream
confluence
0.1
E .
 .
0.05
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 49: Water surface elevation at three control points in Cedar River.
100
x Simulated discharge
x ADCP measured discharge
x ADCP estimated total discharge
50 
E \*
S50
100
150
137 137.1 137.2 137.3 137.4 137.5 137.6 137.7 137.8 137.9 138
Time, days
Figure 410: Measured and simulated discharges through the Cedar River crosssection.
E 20
80
60
80 LLLLLLLL
137 137.1 137.2 137.3 137.4 137.5 137.6 137.7 137.8 137.9 138
Time, days
Figure 411: Measured and simulated discharges through the Ortega River crosssection
(north crosssection of the confluence).
80 r rr
Simulated discharge
ADCP measured discharge
60 x ADCP estimated total discharge
20
E 0
2O
20 *
40
60
80
100 L
137 137.1 137.2 137.3 137.4 137.5 137.6 137.7 137.8 137.9 138
Time, days
Figure 412: Measured and simulated discharges through the Ortega River crosssection
(south crosssection 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/l, 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 (512 g/l), than the values, ranging between 8 and 57 mg/1
(with an additional, exceptional value of 101 mg/1 in one case), obtained from water
sampling on May 17, 2001. Depthaveraged TSS concentration series simulated by the
coarse grid model at the Cedar River crosssection near the confluence is shown in
Figure 413 and commensurate collected samples at the same location are shown in
Figure 414.
Sediment concentration at the Cedar River crosssection 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 413: Depth averaged TSS concentrations at the Cedar River crosssection near
the Cedar/Ortega confluence simulated by the coarse grid model (May 17, 2001).
137.5 137.55 137.6 137.65 137.7 137.75
Time, days
137.8 137.85 137.9
Figure 414: Depth averaged TSS concentrations from the water sample data, collected at
the Cedar River crosssection near the Cedar/Ortega confluence (May 17, 2001).
*
*
*
*
*
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 (89 mg/1),
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/1 in the top layer to 27 g/1 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 415. 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.
x104
Measured data
Linear fit
*
0.8
Z*
S0.6 
0
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 415: 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
c D
concentration oc cannot exceed unity, by definition, hence the constraint
p D1,
D <: D is imposed on the calculation of the diameter.
c
To calibrate the model against the laboratory data (Gowland, 2002) the value of
G = 102 Hz was used, in order to satisfy the above constraint and, at the same time, to
simulate a nearquiescent 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 (230), 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, 2001) was found to be within 8 to 57 mg/1; however for the
modeling purposes the range of concentrations from 0.1 to 100 was selected for
convenience. Figure416 thus obtained qualitatively resembles the Dyer diagram
(Figure 416). The dissipation parameter was related to the shear stress using equation (2
30) considering the mean flow of 10 cm/s.
E 200
50
0
0.8 E 102
2 1 1
Shear stress N/m2 10 Concentration mg/
Figure 416: Calculated floc size as a function of shear stress and concentration.
Figure 417 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
aC"
W = a(4.1)
(b2 +C2)
with the parameters a, b, n and m set to 0.035, 2.0, 3.5 and 2.75, respectively.
W 
E 10
I 10
0)
101 100 101 102
Concentration, kg/m3
Figure 417: 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 (226)],
was taken as 2.3. By fitting the settling velocity predicted by the model to the curve given
by Figure 417, the parameters (k,,k,, 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,w = 1,020 kg/m3 and v = 106 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 (1 and 2) were changed
(Figure 46). 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 41.
Table 41: TSS removal efficiencies of treatment systems in Florida (after Harper, 1997)
Treatment system Estimated TSS removal efficiency (%)
Dry Retention 6098
OffLine Retention/Detention 90
Wet Retention 85
Wet Detention 85
Wet Detention with Filtration 98
Dry Detention 70
Dry Detention with Filtration 6070
Alum Treatment 90
For that purpose, the resulting (calculated) settling flux (total mass of sediment
passing the crosssection 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 predefined percentage. The channel crosssection, 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 46),
each for the selected four removal ratios (0%, 30%, 60%, and 90%). The model was run
for three days, during May 1618, 2001. Four output control points (CP1CP4) were
selected (Figure 46). CP1CP3 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 semidiurnal tidal
cycles (the second cycle on May 17 and two cycles on May 18, 2001) are presented in
Table 42. As seen in this table, it can be inferred that Sites 1 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 42: Comparison of sites with different removal efficiencies with a notrapping
scenario.
Trap efficiencies (%) Net sediment flux g/s Resulting efficiency at the
confluence (%)
Trap 1 Trap 2 Trap 3 CPI CP2 CP3 CP4 confluence (%)
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 6.7 8.8 11.2 5.8 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 43, 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 onetime dredging of the
confluence area to remove soft sediment deposit there.
Table 43: Summary of the effect of treatment on sediment load in the confluence area
Upstream Midstream Downstream Load reduction at the confluence (%)
Cedar Cedar Cedar
Treatment 2
Treatment 13
Treatment 25
Treatment Treatment Treatment 51
CHAPTER 5
CONCLUSIONS
5.1 Summary
A study of the effectiveness of fine sediment trapping in the Cedar River estuary
in north Florida was carried out. A combined threedimensional hydrodynamic and
sediment model was set up and calibrated for this estuary. The original model (EFDC)
was improved for fine sediment settling velocity calculations, by accounting for floc
growth and breakup 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 nearquiescent 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 nearquiescent 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 nearbank 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
(Q,.e,,mre) and estimated (Q,, ) values:
Q1oa, = e, + Qe.a.red (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, = Q,,, + ,,,,,, + Q,,,,, + Qle,, (A. 2)
For calculating the top and bottom discharges two (userdefinable) 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 powerlaw velocity (u) profile in the vertical
(z) direction:
u = r.z (A. 3)
in which the exponent p is userdefined with a default value of 1/6, and the
proportionality constant r is found by fitting the powerlaw profile to the measured
points.
For calculating the right/left discharge the following formula is used:
Q=cDHu,,, (A. 4)
where c is a userdefined 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 userdefined number of leftmost/rightmost ensembles.
APPENDIX B
SETTLING VELOCITY AND FLOC SIZE CALCULATIONS
B.1 Introduction
In order to demonstrate the application of the settling velocity model described in
Section 24 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 (229), (236) and floc aggregation coefficient in a form similar to
C"
equation (41), i.e., kA, =k bC2 The parameters k,,,k,, p and q were determined
(b+C2)
in this way; p and q were found to be 0.6 and 0.45, respectively, and. k, =8.7 and
k,=19.1.103 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 =102 Hz was used to represent
quiescent water and the corresponding velocity curve was plotted (Figure Bl).
In Figure Bl, 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).
/*
Lab results (quiescent water)
x x
. , X xx .
X
Lab results
(Max. velocity=0.09 mis)
1031 ; , : ; I , I
101 100 101 102
Concentration g/l
Figure Bl: Settling velocity calculation test results, and comparison with data of
Wolanski et al. (1992) using sediment from Townsville Harbor, Australia.
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 EmsDollard River area in The
Netherlands. Particle sizes were measured using a Malvern particle sizer. For simulation
purposes the parameter values in Table Bl, plus those provided by Winterwerp, were
used. The initial particle size was taken as 4 pm, as measured by Winterwerp.
Coefficients k,, = 14.7 and k, = 14.0 103 were selected.
Comparisons between simulations and data are shown on Figure B2. The values
of concentration and dissipation parameter used are given in Table B1. The resulting
curves, which lead to equilibrium sizes, appear to be the same as those of Winterwerp
(1998). Floe 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 I
10 10 10 10"
Time, s
Figure B2: Floc growth with time measured and predicted for River EmsDollard mud
(Winterwerp, 1998).
Table B1: Data from settling column tests with EmsDollard mud.
c G
Test No C G
TestNo (kg/m3) (Hz)
T73 1.21 81.7
T69 1.17 28.9
REFERENCES
Dyer, K. R., 1989. Sediment processes in estuaries: Future research requirement. Journal
of Geophysical Research, 94(C 10), 94899498.
Ganju, N. K. 2001. Trapping organicrich 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/COELCR/2002/001,
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, 619.
Huang, H., 1994. Fractal properties of flocs formed by fluid shear and differential
settling, Physics of Fluids, 6(10), 32293234.
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), 883892.
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, 1940.
Mellor, G. L., and Yamada, T., 1982. Development of a turbulence closure model for
geophysical fluid problems. Reviews in Geophysics and Space Physics, 20, 851
875.
National Research Council, 2001. A riskmanagement strategy for PCBcontaminated
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, 479486.
Stoddard, D. M., 2001. Evaluation of trap efficiency in an estuarine environment. Report
MRP2001/003, 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), 469483.
Teeter, A. M., 2001. Claysilt 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, 157170.
Van Leussen, W., 1994. Estuarine macroflocs and their role in finegrained 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), 309326.
Wolanski, E., Gibbs, R., Ridd, P., Mehta A., 1992. Settling of oceandumped dredged
material, Townsville, Australia. Estuarine, Coastal and Shelf Science, 35, 473
489.
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.
