COHESIVE SEDIMENTS IN COASTAL ENGINEERING
Ashish J. Mehta
COHESIVE SEDIMENTS IN COASTAL ENGINEERING APPLICATION1
Ashish J. Mehta
Coastal and Oceanographic Engineering Department
University of Florida, Gainesville, FL 32611
I am honored by the opportunity provided by TOA Corporation
to present to the engineering and the scientific community of Japan
a few salient aspects pertaining to cohesive sediment transport in
coastal engineering applications. Such applications, as related to
port development, dredging of approach channels, coastal turbidity
generation, erosion or accretion of mudflats and muddy beaches,
contaminant transport associated with fine sediment movement etc.,
require a knowledge of the basic transport principles and
methodologies for measurement and analysis within a prescribed
technical framework. At the University of Florida we have
participated in this type of development over the past couple of
decades; hence the following presentation focuses on a perspective
which primarily reflects our own experience. Needless to say other
compatible perspectives do exist.
It is most essential to recognize that there are general
similarities as well as significant dissimilarities between fine-
grained cohesive sediment transport and coarse-grained transport
such as of sand. Consider for example the dumping of dredged
material from a hopper in the offshore area. As depicted in Figure
1, the short term outcome usually is the development of near-bed
fluid mud layers and a high concentration turbidity current which
may spread to significant distances depending on the initial
(negative) buoyancy of the dumped slurry, and the intensity of wave
and current action. A settled bed will form only after the near-
1Based on lecture presented in Tokyo on November 30, 1989,
sponsored by TOA Corporation, Yokohama.
bed material has had enough time to dewater. This can be contrasted
with the dumping of sand which will tend to form a bed rapidly,
thereby capping the underlying substrate. In fact, as a result of
the low settling velocities of fine sediment, and low permeability
of the bottom deposit compared to sand, the response times related
to transport and settling in the former case tend to be
considerably larger than in the latter.
Another feature of cohesive sediments which is characteristic
of their fine particle size is their high adsorptive capacity for
nutrients and contaminants. A typical particle is platy (Figure
2), with a very large surface area per unit weight. This surface
provides a large number of adsorption sites for foreign atoms and
molecules. Such individual particles agglomerate in saline waters
(salinity exceeding 2 to 3 ppt) to form aggregates consisting of
a very large number of individual particles. These agglomerates are
often covered by biogenic detritus thereby further enhancing their
sorptive capacity. In Figure 3 the concentrations of four sorbed
trace metals are plotted against the percent of sedimentary
material of size less than 16 microns from the work of Salomons and
Mook (1977). Increasing percentage implies decreasing particle
size. As observed the concentrations of zinc, chromium, lead and
copper increase consistently with percentage. In general therefore
contaminants go with the fine component of the sediment load and
degrade water quality at depositional sites. Furthermore, the
presence of reducing agents can create a highly anoxic environment
in the benthic boundary layer and destroy marine life forms.
In what follows we will consider the transport properties of
agglomerated fine sediments followed by some illustrations of
application methodologies and case studies.
AN APPLICATION PERSPECTIVE
As engineers and scientists we are usually called upon to
investigate a macroscale problem in relation to the microscale
processes which ultimately lead to such a problem. In Figure 4(a)
the linkage between macroscale coastal topographic changes and the
corresponding microscale near-bed processes is shown in a
descriptive manner. The typical causative factor for problem
occurrence is anthropogenic, e.g. dredging or shoreline
modification due to port construction such as at Kumamoto (Japan
International Cooperation Agency, 1988), which influences the
topography directly, and also affects hydrodynamic forcing. Both
in turn determine water column dynamics which is linked to muddy
bed dynamics via flow-bottom interaction. Understanding flow-
bottom interaction and the manner in which it links bed response
to water column dynamics constitutes the core problem for cohesive
sediment transport investigation.
A common illustrative problem is the prediction of the rate
of accumulation of "fluff" at the channel bottom in order to
estimate the frequency of maintenance dredging (Figure 5). The
highly skewed nature of the concentration (C) profile, which is
heavily bottom weighted, coupled with turbulent flow velocity (u)
profile, which may itself be influenced by the concentration field,
can lead to residual sediment flux largely confined to the near-
bed zone. Important points to be made are that: 1) the vertical
structures of the interactive concentration and velocity fields
govern the horizontal transport rates, 2) suitable measuring
instrumentation must be deployed to account for the motion of the
near-bed sediment slurry which may constitute the overwhelming
component of the total rate of mass transport, 3) since the rate
of residual transport is important in determining the rate of
channel infilling, accuracy in calculating the transport rate in
a given direction must be high, and, therefore, the constitutive
properties of the slurry, which dictate transport in conjunction
with hydrodynamic forcing, must be carefully identified, and 4)
given that the bottom response will be predominantly tide
controlled in macrotidal areas but largely episodically determined
in microtidal and wave dominated environments, the time scales of
interest can vary from days to months to years.
A somewhat different perspective results if boxes (A) and (B)
in Figure 4(a) are replaced, for example, by corresponding ones in
Figure 4(b). In the event the problem becomes one of water quality,
in particular that of eutrophication. Increased trophic levels
manifesting in algae bloom due to external loading by phosphorus
is a common problem in lakes. Although the water environ is clearly
different from that of the coast, the bottom sediment resuspension
and deposition mechanisms responsible for the release of soluble
reactive phosphorus due to wave action in the lake are basically
the same as those that occur due to coastal waves.
The relationship between soluble reactive phosphorus and the
resulting biomass is schematized in Figure 6 (Lijklema et al.,
1983). Omitting details, the matter of interest here is that
related to sediment resuspension, which consequently results is the
desorption of suspended sediment bound phosphorus. Resuspension in
this context is essentially inclusive of sediment settling (as
distinct from biomass settling) and deposition. Response time
scales of typical interest range from a few hours to a few days
essentially tied to weather fronts. These time scales are therefore
typically shorter than those associated with channel infilling for
example. Once again however, despite the dominance of convective
transport over advection, the constitutive properties of the bottom
mud play a critical role in governing suspended sediment transport.
The normal mode of sediment classification by particle size
suffers from some limitations which must be recognized at the
outset. In Figure 7 the dispersed particle size distributions of
three sediments are given for illustrative purposes (Mehta, 1973).
These were obtained by the standard hydrometer test. It should be
noted that while silt is separated from sand at 60 microns,
cohesion becomes important only for sizes less than around 20
microns. Thus, the 20 to 40 micron range marks the transition
between cohesionless and cohesive particles. Clay minerals, which
are by definition smaller than 2 microns, constitute the main
cohesive ingredient of muds. Kaolinite, illite and smectite
(montmorillonite), all composed of silica, alumina and water, but
having different crystalline structures, are the three principal
clay minerals. Kaolinite is the least cohesive of the three while
smectite possesses the highest degree of cohesion, as reflected by
the magnitude of the cation exchange capacity (Grim, 1968). The
Maracaibo mud in Figure 7 had a fair amount of silt and some fine
sand, and was the least cohesive of the three sediments indicated.
San Francisco Bay mud was more cohesive than the kaolinite despite
larger particle size in Bay mud, because it had illite and smectite
as the principal constituents, besides kaolinite.
Since sediments larger than 60 microns are considered to be
coarse while those less than this size fine-grained, in conjunction
with the aforementioned comments, particles between 40 and 60
microns are fine-grained but largely cohesionless. Hangzhou Bay in
China is an example of a coastal bay in which the bottom sediment
is mainly of this type (Costa, 1989).
In colloidal chemistry a fluid-particle mixture consisting of
particles smaller than 1 micron is defined as a sol, while a
suspension consists of larger particles (van Olphen, 1963).
Particles in a sol remain perpetually in suspension due to Brownian
motion. As will be seen subsequently however, the definition of a
cohesive sediment suspension from the transport perspective must
be based on different physical principles.
The significance of cohesion in binding very small particles
can be simply illustrated by calculating the electrochemical force
relative to particle weight. For example, the weight of a particle
of 1 micron nominal diameter is 2.6x10-14 N. We next need a measure
of the cohesive force. We know from theological experiments on
flocculated suspensions (Krone, 1963) and erosion of floc layers
(Parchure and Mehta, 1985; Mimura, 1989) that the bulk shear
strength of the individual floc is on the order of 0.1 N/m2. We
will assume, for the sake of present argument, that this strength
reflects the particle-integrated cohesive bond strength; hence the
corresponding force per particle would be 1x10-13 N. This in turns
means that the cohesive force is an order greater in magnitude than
gravity. In actuality, interparticle electrochemical attraction is
due the omnipresent London-van der Waals force, which for example
causes fine particles to stick to the walls of a glass container,
forming a thin film over the surface.
As with cohesionless sediment transport, the settling velocity
of the aggregate is the most fundamental property associated with
cohesive sediment transport. As a result of aggregation due to
interparticle collision and mutual particle interference, the
settling velocity of cohesive aggregates varies strongly with
suspension concentration (volume fraction). Consequently the mass
settling flux also varies nonlinearly with concentration.
Aggregation occurs by virtue of net attractive forces between
particles, brought close enough by Brownian motion, differential
settling and flow shearing. Although the relative importance of
collision frequency due to these mechanisms depends on the
particle diameter, shearing under turbulent flows seems to be the
most important factor contributing to aggregation leading to strong
bonds, with the exception of slack water periods when differential
settling becomes dominant (Krone, 1962). Aggregates or flocs are
formed of individual particles and can, themselves, form aggregates
of higher orders (e.g. Figure 2). They differ from primary
particles in four main aspects: 1) their size is orders larger
than that of individual particles- macroflocs can attain a size on
the order of millimeters (Dyer, 1989); 2) their density is
considerably less than that of the particles due to interstitial
water; 3) their shape is more spherical than the plate-like shape
of the primary particles, leading to reduced drag; and 4) they are
extremely weak, tending to break up easily (Lick, 1982).
Among the above factors, the effect of increase in fall
diameter relative to primary particle is much more significant than
the decrease in density, with the result that the settling
velocities of aggregates are substantially higher than those of
individual particles. The magnitude of the aggregate diameter and
settling velocity are, moreover, only slightly dependent on the
primary particle diameter. The dependence of the settling velocity
of cohesive sediment particles (primary particles or flocs) on the
concentration, neglecting usually secondary effects such as those
of temperature and salinity, falls within three regimes, as shown
in Figure 8(a).
Free Settling: Free settling occurs at low concentrations,
typically lower than C1 = 100 to 300 mg/l (Krone, 1962; Mehta,
1988). Here, concentration has been conveniently expressed as dry
sediment mass per unit volume of suspension, rather than as volume
fraction. In this range the particles settle without mutual
interference; their terminal settling velocity being the result of
a force balance between drag and net negative buoyancy. The
settling velocity, ws, is given by
w r2gD (Ps 1 /2(1)
s Co PW
where D is the particle diameter, ps and Pw are the grain and fluid
densities, g is acceleration due to gravity and CD the drag
coefficient. Fine sediment in dispersed or quiescent condition
typically falls within this regime, although the generally non-
spherical shape of the particles requires the use of an effective
particle diameter, and the settling velocity is modified thereby,
particularly as the particles become very small and graded (Lick,
1982; Ross, 1988).
Flocculation Settling: When the suspension concentration
becomes higher than the free settling limit, increased
concentration and interparticle collision cause an increase in
aggregation and higher settling velocities. The general expression
for ws in the flocculation settling range is
ws = klC (2)
The coefficients in Equation 3 may be determined in laboratory
settling columns or in field tests; values determined by the latter
procedure have sometimes been found to be higher by as much as an
order of magnitude than those in the laboratory, using the same
sediment. This is mainly due to the effect of local flow shearing
rate on kl. On the other hand nl, which is typically equal to 4/3,
may change somewhat as well depending upon the level of turbulence.
In situ measurement using an "Owen tube" is one way to obviate the
problem of scaling laboratory data for prototype application (Burt,
Hindered Settling: For concentrations higher than a value C2
of about 2 to 5 g/l, the settling velocity decreases with
increasing concentration. This is the result of hindered settling,
a phenomenon in which the aggregates become so closely packed that
the fluid is forced to flow between them through increasingly
smaller pores (Kynch, 1952). The general expression for the
settling velocity in the hindered settling range is
ws = ws0[1 k2(C C2) (3)
where Ws0 is the value of ws at C = C2, k2 is the inverse of the
concentration when ws = 0 and, theoretically, for falling spheres
n2 = 4.66 at low Reynolds numbers (Richardson and Jeronimo, 1979).
In general, wso, k2 and n2 are required to be determined in
settling columns (Ross, 1988). Cohesionless sediments also exhibit
hindered settling behavior, but obviously not flocculation
settling. Therefore, the experimentally determined value of n1
(Equation 2) can be used as an indicator of the degree of cohesion
of the sediment (Figure 8(b)).
Fig. 8(c) shows an example of settling velocity data using
mud from Lake Okeechobee, Florida obtained in a 2 m tall settling
column (Hwang, 1989). The relationship
S= 2 C (4)
ws (C2 + b (4)
modified from Wolanski et al. (1989), has been used to represent
both flocculation settling and hindered settling, with a = 33.38,
b = 2.54, n = 1.83 and m = 1.89. Note that the corresponding curve
for the flux, Fs = wsC, is obtained by replacing n by n + 1 in
Equation 4. Note further that the peak value of w. is defined by
n-2m n- i) 2
so ab2m (5)
C2 = 1/2 (6)
Fishermen as well as coastal engineers are aware of the
tremendous role played by mud flats in dissipating the energy of
propagating waves as they approach the beach. Understanding the
energy absorbing role of soft coastal muds requires a knowledge of
the constitutive behavior of mud as obtained through rheometric
measurements. Models have been developed to calculate the degree
of wave damping assuming non-Newtonian constitutive properties,
e.g. pseudoplastic (Figure 9) or viscoplastic. Similar models have
also been used to calculate the horizontal mass transport of mud
resulting from the nonlinear interaction between waves and soft
mud beds (Sakakiyama and Bijker, 1989).
Recent works on mud rheology point of deficiencies in
traditional approaches to rheometry for obtaining answers usually
sought in mud related investigations (Williams and Williams,
1989a,b). Furthermore, these works clearly lay out a case for using
a constant stress (as opposed to constant rate of strain)
rheometer, as well as a pulse shearometer, not only for determining
the viscosity and the modulus of elasticity, but also for
identifying a yield strength that truly defines the stress beyond
which the bed structure loses its integrity, and for characterizing
the rigidity of the cohesive bed.
In general, the interaction between waves and muddy substrates
is quite complex, particularly because wave action itself causes
time-dependent changes in the bottom mud properties leading to its
degradation over time. Erosion may occur if and when the bed
becomes very weak. This is illustrated schematically in Figure 10.
The description is largely self-explanatory, but for details
reference may be made to Maa and Mehta (1989). A sophisticated
numerical model using essentially the same concept has recently
been developed and applied by the Port and Harbor Research
Institute to determine the height of submerged protective levees
to prevent sedimentation in the navigation channel approaching the
new port of Kumamoto (Tsuruya et al., 1990).
Vertical Structure of Concentration
In accordance with the fact that the rate of horizontal
sediment transport is governed by the dynamics of the vertical
variation of sediment concentration, a definition-related
description for the same is presented in Figure 11(a) (Mehta,
1989b). The top mixed layer of mobile suspension is defined as one
which is low in concentration, not exceeding about 0.3 g/l. This
suspension is practically Newtonian, and within it the fine
sediment aggregates settle more or less freely, without significant
interparticle interference. Upward diffusion is neutral, largely
uninfluenced by sediment. Below this layer settling as well as
diffusion are conditioned by sediment concentration effects, and
stratification begins to become apparent. The settling velocity
increases with concentration due to flocculation resulting from
interparticle collision, and the vertical diffusivity is reduced
due comparatively large concentration gradients. A step-like
microstructure defined by these gradients, or lutoclines, is
characteristic of this regime in which the suspension becomes
increasingly non-Newtonian, typically pseudoplastic at low shearing
rates with increasing concentration.
The lutocline shear layer is perhaps the most significant
feature of the suspension profile, formed as a result of buoyancy
stabilization of the high concentration suspension. Below this
layer the sediment settling flux is hindered at concentrations
exceeding anywhere between about 4 and 20 g/l. This concentration
gradient, which is itself determined mainly by hindered settling
effects, may be called the primary lutocline, as opposed to the
secondary lutoclines above whose structure is governed by settling
combined with upward diffusion (Ross and Mehta, 1989; Scarlatos and
Mehta, 1990). Turbulent boundary layer effects from the fluid
column above result in shear production and sediment entrainment
at the lutocline; however, from the point of view of the vertical
distribution of sediment mass the gradient tends to remain largely
intact even in high energy situations, although its elevation may
change somewhat with time (Kirby, 1986; Ross and Mehta, 1989).
The mobile, hyperpycnal layer commonly called fluid mud can
occur up to concentrations on the order of 200 g/l. It is
essentially a fluid supported slurry which in some instances
approaches Bingham plastic behavior, and plays a significant role
in absorbing and dissipating turbulent kinetic energy. In addition
to upward entrainment or downward mobility resulting from
dewatering, fluid mud can move horizontally via an applied stress
at the elevation of the lutocline, or down gentle natural slopes,
thus contributing measurably to mass transport. The stationary mud
layer below the fluid mud may or may not be wholely fluid supported
depending upon its stress history. If measurable effective stress
develops due to consolidation, then this layer is defined as a
cohesive bed. Stationarity is here defined in reference to zero net
horizontal motion; oscillatory motion is not precluded and in fact
may be quite significant under storm wave action. Waves, via
transmission of normal and shear forces, build up excess pore
pressures which essentially inhibit the structural integrity of the
porous bed matrix. Wave particle orbits can as well penetrate the
cohesive bed causing it to undergo viscoelastic deformation, which
may eventually lead to its fluidization.
An evident consequence concentration (C(z)) weighting towards
the bottom is that the horizontal mass flux of sediment (uC(z)) is
likewise weighted as well, as depicted in Figure 11(b). This
condition makes it imperative, 1) to device means to identify the
zero velocity level and its time-variation in the field, and 2) to
accurately measure both the velocity (u(z)) and the concentration
profiles above this level. Both theory and requisite
instrumentation for this purpose are currently being developed. As
indicated in the figure, the sediment mass transport velocity can
be as low as one-half the mean current.
Inherent in the description of Figure 11(a) is a change from
a water-sediment mixture state at low concentrations to a fluid
state at high concentrations (Sills and Elder, 1986). Furthermore,
with increasing concentration, turbulence is damped, and at very
high concentrations the flow may become essentially viscous. As
implied in Figure 12, under low concentration, turbulent flow
conditions the particulate behavior of the mixture is apparent,
while at the other extreme, in the hyperconcentrated state and at
concentrations even higher, the dynamical properties of the soil
fabric become important. At somewhat lower concentrations the mud
exhibits essentially a non-Newtonian viscous fluid behavior.
Finally there exists a wide range of flow and concentration defined
conditions over which particulate as well as fluid behaviors tend
to coexist. A real challenge for the future is to develop theories
to explain the complex dynamic response that results from this
combination of behaviors.
The dynamic response of the concentration profile is of course
quite sensitive to specification of the vertical sediment mass
fluxes. With regard to settling, the highly nonlinear dependence
of the settling velocity on concentration of the suspended solids
(Figure 8(a)) must be specified for sediment under consideration.
Specification of diffusion must recognize damping due to
stratification (Ross and Mehta, 1989). Damping essentially results
in a nonlinear dependence of the diffusive flux on the
Commonly used, bed shear stress based relationships for
erosion and deposition (Parchure and Mehta, 1985; Umita et al.,
1984; Mehta and Lott, 1987; Mehta et al., 1989a) are generally
specified at the elevation where the stationary cohesive bed is
encountered. As noted, however, consideration of bed deformation
requires specification of the energy dissipative properties of the
mud. Thus, for example, a viscoelastic, Voigt element description
can be used for the mud constitutive behavior to solve for the
oscillatory mud motion (Maa and Mehta, 1987). In comparatively low
energy, low concentration situations deposition rates are as well
sufficiently low such that the deposit dewaters rapidly and forms
a settled bed. In this case therefore the total thickness of the
layer comprising the stationary mud, the mobile hyperpycnal layer
and the lutocline shear layer may be practically negligible.
Consequently the closely Newtonian mixed layer mobile suspension
together with the cohesive bed specify the system adequately, and
the usual stress based erosion and deposition rate relationships
yield sufficiently accurate description of the mixed layer
concentration profile. As suspension concentration increases
however, these rate relationships become less representative, and
in high concentration environments take on primarily a qualitative
role (Costa, 1989). As depicted in Figure 13, in the event the
processes that need attention are settling, bed formation,
consolidation, and gelling fluidization and entrainment of
fluidized mud (Mehta, 1989a).
The limitation of bed shear stress based erosion flux modeling
for fluid mud, which by definition can not resist deformation, has
been well recognized previously (Parker and Kirby, 1982). In fact,
the analogy between salt- and sediment-induced stratification in
the context of fluid mud entrainment, in contrast to bed erosion,
is rather evident (Odd and Cooper,1989). Quantification of the rate
of entrainment of a fluidized mud layer underlying a flowing mixed
layer, requires an understanding of interfacial dynamics. The
characteristic governing parameter, namely the Richardson number,
Riu, can be selected to be equal to hAb/u, where u is the mean
mixed layer velocity, h is the mixed layer depth, and Ab is the
interfacial buoyancy step characterized by b= g(p Pw)/pw, with
p and Pw the density of the sediment-laden fluid and water density,
respectively. Entrainment proceeds via generation of interfacial
instabilities and mixing. The relationship between the non-
dimensional entrainment flux, Q= q/(uAb), and Riu, where q=
(g/pw)dm/dt; dm/dt being the upward sediment mass flux, conforms
approximately to the expression (Srinivas, 1989)
Q = A Ri-0.9/[B2 + Ri ] (7)
where A= 0.27, B= 20 and m= 0.66 are values of the free
coefficients obtained by Srinivas in his specific experimental
setup. Beyond Riu 25 the intensity of entrainment can be highly
diminished. Apart from the fact that Equation 7 better represents
the mechanics of entrainment than the stress-based erosion models,
the nature of this essentially energetic based relationship
implies the dependence of dm/dt on u3, as opposed to u2. Evidently
therefore fluid mud entrainment follows a different law than that
for bed erosion. This difference may be particularly important when
strong, episodic responses are involved.
A measure of how well a given bottom erosion flux relationship
predicts suspended sediment behavior is obtained by looking at the
well known hysteresis relationship between concentration and mean
flow velocity almost always found in tide dominated systems. In
fact, prediction of flow-sediment hysteresis in coastal water
bodies is tantamount to the recognition of the role of tidal
pumping in governing macroscale turbidity transport phenomena
(Dyer, 1986; Futawari et al., 1988). It is evident that, for
instance, in the high concentration environment the rate at which
bottom material is suspended and the rate at which the suspension
settles out are extremely sensitive to the physical state of the
material, hence proper representation of this state is quite
It is feasible to examine the role of bottom fluxes in the
above context by carrying out simple, 1-D vertical model simulation
of hysteresis loops. This is illustrated here based on data
obtained at a nearshore site in Hangzhou Bay, China, a mesotidal
coastal waterbody consisting of a silty mud substrate (Costa,
1989). Figure 14 shows the simulated variation of (turbulence-
mean) sediment concentration, C, with the square of the
corresponding horizontal velocity, u (plotted as ulul) at two
elevations, 1.25 and 2.75 m, above the bed over one tidal cycle.
Elapsed times in minutes are indicated against the computed points.
The stress based (proportional to ulul) erosion and deposition rate
relationships including settling velocities required for this
simulation were obtained in laboratory experiments using bottom
collected sediment. Comparing these loops with measured ones,
although incomplete, in Figure 15 indicates that the observed
features have been adequately simulated in a general sense. It was
found however that, in order to achieve order of magnitude
agreement between the measured and simulated loops, while the
settling velocity data from the laboratory seemed to be adequate
for prototype loop simulation for this particular sediment (which
was fine-grained and weakly cohesive), significant adjustments had
to be made in specifying the "free" coefficients characterizing the
bed fluxes and the vertical mass diffusivity. In particular, the
bed erosion shear strength and the critical shear stress for
deposition (Mehta, 1988) had to be assigned values which were an
order larger than those measured under laboratory conditions.
Likewise, the Richardson number dependent mass diffusivity used
(order of 0.1 m2/s maximum) was considerably greater than measured
(order of 0.001 m2/s). It is also interesting to note that the
measured turbulent Schmidt number ranged from 0.94 to 2.40, which
highlights the need to use mass flux-based rather than velocity
gradient-based Richardson number in modeling diffusion (Oduyemi,
Based on the types of difficulties in hysteresis loop
simulation, it can be concluded that simple, stress based modeling
of vertical transport can not fully account for the complexities
in bottom mud structure and dynamic response to tidal current
forcing (Costa, 1989). Furthermore, calculation of the bed shear
stress from turbulence-mean flow velocity also introduces
limitations in accounting for the rate of sediment entrainment,
partly due to hysteresis between this velocity and related
turbulence dependent quantities (Dyer, 1986). One representation
of this hysteresis which, to a measure, is influenced by the non-
equilibrium between the rates of turbulence production and
dissipation in the estuary, is shown in Figure 16 in terms of the
relationship between the measured Reynolds stress and the velocity,
u, 1.25 m above the bed.
PROBLEM SOLUTION STRATEGY
Simulation of fine sediment transport for verification as well
as predictive purposes can range from empirical to fully three
dimensional numerical modeling. It is self-evident that while the
development of solution options must rely on the outcome of the
result of simulation, the latter in turn must be firmly grounded
in physical evidence from the field and associated laboratory
experiments. Among the available problem solving approaches, the
so called hybrid approach is technically obviously meritorious in
that it carefully combines the strengths of physical and
mathematical modeling. A comprehensive methodology of this type
has been developed over a number of years at the U.S. Army Engineer
Waterways Experiment Station, Vicksburg, Mississippi (McAnally et
Some illustrations of application methodologies are very
briefly mentioned in what follows, beginning with the problem of
predicting the future infilling of a coastal bay.
Bay Infilling by Fine Sediment
Prediction of sedimentation in bays can be illustrated by the
case of Atchafalaya Bay in Louisiana (Figure 17). The Atchafalaya
River, a distributary of the Mississippi River, discharges into
this bay. In recent years, the delta at the mouth of the river has
grown dramatically. An investigation of the bay and adjacent waters
was carried out to predict the rate at which the delta will evolve
in the short term (- 10 years) and the long term (50 years), and
the manner by which that evolution will affect flood stages,
navigation channel shoaling, and the environmental resources of the
area (McAnally et al., 1984b; Mehta et al., 1989b).
A number of factors combined to make the study unusually
complex. They included the long period over which predictions had
to be made; the migration of the region of delta growth from
lacustrine to estuarial to marine environments; a hydrodynamic
regime that is variously dominated by river flows, wind-induced
currents, tides, waves and storm surges; and the combined
deposition of sediments from the sand, silt and clay classes. In
order to adequately address this problem, the investigation
included several predictive techniques, including: 1) Extrapolation
of observed bathymetric changes into the future; 2) a generic
analysis to predict future delta growth by constructing an analogy
between the Atchafalaya delta and other deltas in similar
environments: 3) quasi-two-dimensional numerical modeling of
hydrodynamics and sedimentation; and 4) use of extensive field and
Results showed a wide range of possible future land growth
rates for 50 years in Atchafalaya Bay, and highlighted the
sensitivity of delta growth to bottom subsidence rate. High
subsidence rates there are caused in part by compaction of thick
layers of fine sediments that have been deposited by the
Mississippi River and its distributaries over millennia.
Delta extent (> -3 ft) in the year 2030 shown in Figure 18(a)
using the extrapolation approach was generated by establishing a
relationship between past delta growth and forcing phenomena of
river flow and sediment supply, then using that relationship in
combination with historically recorded flows to project future
delta growth. It did not explicitly include subsidence effects. The
quasi-two-dimensional approach (Figure 18(b)) employed a one-
dimensional model with a subsidence rate of 1 cm/yr. Even without
any provision in the model to account for redistribution of bottom
sediment by resuspension due to currents and waves, delta growth
is predicted to be considerably lower than in Figure 18(a). Recent
observations in the bay suggest that the rate of emergence of the
bottom is measurably lower than that suggested by the extrapolation
technique (McAnally, personal communication). This outcome points
to the critical importance of seemingly second order effects, e.g.
subsidence, in influencing the evolution of sediment-related
Sedimentation in Closed-End Canals
Closed-end channels such as pier slips, tidal docks, elongated
marinas and residential canals are well known sites for fine-
grained sediment deposition. For example, in an investigation of
pier slips at the Mare Island Naval Shipyard in the San Francisco
Bay area, Jenkins et al. (1980) found the rate of sedimentation in
the slips to be as much as 2.5 times higher than that in the main
channel to which the slips were connected. Similar instances are
also found at a large number of ports worldwide.
In general, advective transport due to tidal prism, wind-
driven circulation and density-induced currents are processes by
which sediment enters the closed-end channel. Turbidity current is
driven by the difference in density between the sediment-laden
outside waters and the relatively quiescent and sediment-free
waters in the channel. The contribution to the total rate of
sedimentation from this mechanism evidently varies with the
physical conditions. In areas where tides are weak or when the
suspended sediment concentration is high, turbidity current becomes
an important source of sediment in the channel. For example, in
many residential canals in Florida, sedimentation largely occurs
during storms when the suspended sediment concentration in the
waterways increases by one to three orders of magnitude over that
during fair weather (Maa and Mehta, 1985). Sediment influx under
these episodic conditions can be thought of as occurring in bursts
due to opening of a hypothetical gate (Figure 19); these bursts
being separated in time by calms with relatively low level ambient
concentrations (5-10 mg/l).
Figure 20(a) shows the location of Orange Waterway, a
residential canal off Marco River, a microtidal coastal waterbody
near the southwestern coast of Florida. The rate (cm/yr) and
pattern (longitudinal variation) of deposition observed in Figure
20(b) (Wanless, 1975) seem to be largely consistent with laboratory
observations on turbidity transport in closed-end channels (Lin and
The decay of mean concentration below the interface with
distance which accompanies the ingress of the turbid current and
the influence of an important governing parameter, namely the
dimensionless settling velocity, wsl/uA, on the longitudinal
concentration distribution are apparent in Figure 21, in which the
concentration ratio, Cb/Cbl at steady state (when the rates of
inflow and outflow become equal) is plotted against dimensionless
distance, x/H. Note that Cbl is the value of the mean concentration
in the lower layer, Cb, at the mouth and H is the depth of water.
The settling velocity at the mouth, wsl, has been normalized by uA,
the interfacial celerity. In non-stratified turbulent flows, uA is
replaced by the friction velocity. The data were collected in a
laboratory flume (Lin and Mehta, 1989). Results from seven tests
using kaolinite have been included. Three lines represented by
different values of wsl/uA are shown. These lines are based on an
analytic solution for the decay of concentration with distance.
Line slope increases with increasing wsl/uA, since, for a given uA,
increasing wsl implies increasing rate of deposition.
The rate of sediment influx through the mouth, S, can be
practically calculated by dividing the total deposited sediment
mass in the channel by the influx duration and the area of the
lower half of the flow cross-section at the entrance. The quantity,
S, is plotted against entrance concentration C1 (which is
practically equal to Cbl) in Figure 22. A 3/2 power dependence of
S on C1 is evident. This relationship has a rather simple
phenomenological basis (Lin and Mehta, 1989).
Sedimentation in Small Marina Basins
Prediction of the pattern and rate of sedimentation in marinas
is required at the design stage for the estimation of capital and
maintenance costs of dredging, as well as implementation of
measures for minimizing sediment intrusion into the basin. In
general, given basin geometry and sediment load in waters outside
the basin, the nature of flow circulation principally determines
the degree of sedimentation within the basin.
The number of entrances and the planform shape of basins vary
widely. In areas where suspended sediment loads are relatively
high, such as in the San Francisco Bay system, it is desirable to
construct water-tight basin walls and a single entrance in order
to minimize sediment intrusion. An examination of the planform
geometry of single entrance marina basins in Florida revealed that,
on the average, the basins are rectangular of 100 m sides, with a
55 m long entrance channel connecting the basin to the main water
body (Srivastava, 1983). The corresponding mean water depth in the
basin was found to be 2.2 m. In basins of this type, or even in
those with more than one entrance, given the planform shape and
depth, the principal factors which influence the degree of
sedimentation are: 1) the range and period of tide, 2) wind, 3)
waves, 4) salinity gradients and 5) concentration in outside
waters, provided the flow environment in the basin is predominantly
depositional, i.e. a basin in which there is no significant
resuspension of the deposit. If erosion does take place (for
instance when the basin has two entrances of such a configuration
as to cause appropriately strong currents during at least a portion
of the tidal cycle), the erosional properties of the deposit must,
in addition, be known.
Reference must be made here to the nature of suspended
sediment concentration outside the basin. Observations have shown
that in estuaries with large tidal ranges, there is an acceptable
degree of correlation between the depth- and tide-averaged
suspension concentration, CO, the range of tide, and even
temperature, T. For example, measurements in the Mersey Estuary,
England, yielded the following linear relationship (Halliwell and
CO = a + pH + 6T (8)
where H = high water elevation (not the tidal range) above
Liverpool Bay Datum, T = temperature and a, p and 6 are empirical
coefficients. Relationships such as Equation 8, once evaluated for
a given site, provide a convenient means to estimate the temporal
variation of CO at that site, given the corresponding variation of
H, and T (which in many cases can be assumed to be constant, at
least on a seasonal basis).
In episodically controlled environments, however tide-
determined relationships such as Equation 8 are of limited use and,
in general, it becomes essential to obtain extensive measurements
of suspension concentration at the site. Figure 23 shows a
concentration histogram (frequency of occurrence, 4, of the
suspension concentration, CO) derived from measurements near
Camachee Cove Yacht Harbor basin (Figure 24) on Tolomato River near
the Atlantic Coast of Florida (Srivastava, 1983).
In general, prediction of the pattern and rate of
sedimentation in the basin involves numerical solutions of the
advection-dispersion equation for sediment mass transport with
appropriate descriptions of deposition and erosion, given the flow
field. Results obtained in this way usually show a good degree of
agreement with measurements as far as the basin-average rate of
sedimentation is concerned, provided the flow field is correctly
represented. On the other hand, the pattern of deposition is
complicated by the presence of docks and boat traffic, as well as
by redistribution of the sediment during episodic events. Since
these factors are not easily accounted for in numerical modeling,
detailed patterns are often not simulated adequately; however,
areas of critical shoaling, if prominent, are often identified.
For many small marina-type design problems, it is sufficient
to estimate the basin-average shoaling rate without considering
the depositional pattern. Simple computational procedures, even
though far less accurate than sophisticated numerical approaches,
can be quite useful in such cases. In Figure 25, the computed
shoaling or sedimentation rate, SR, on an annual basis is plotted
against the outside concentration, CO, for Camachee Cove basin
using a zero dimensional approach (Maa and Mehta, 1985). Using the
observed variation of CO according to Figure 23, the actual total
rate of shoaling per year may be obtained by utilizing the results
of Figure 25. For each CO corresponding to a frequency 4 from
Figure 23, the SR value is obtained from Figure 25. This SR is then
weighted by multiplying it by the corresponding value of ). The
total rate of shoaling is the sum of '*SR, which in the present
case can be shown to be equal to 0.13 m/year. Bathymetric
measurements in the basin over a 2.5 year period yielded 0.15
Note that in this illustration the settling velocity, ws, was
taken to be proportional to 4/3 power of concentration C according
to Equation 2 (nl = 4/3), thus incorporating the effect of floc
aggregation on deposition. As a consequence the settling flux, wsC,
becomes proportional to C to the 2.33 power. This consideration
leads a higher rate of shoaling than would be the case if ws were
assumed to be constant, and thus yields an answer on the
An example of predicted sedimentation pattern (contours of
deposit thickness per year) in the Camachee Cove basin is shown in
Figure 26. A depth-averaged suspended sediment transport model
was used for this purpose (Hayter and Mehta, 1986). The flow field
necessary for driving the sediment model was generated through a
compatible hydrodynamic model. The predicted basin-average deposit
thickness is 14.6 cm/yr.
Erodible Depth due to Wave Action
As noted, a key feature of fine-grained suspended sediment
concentration profiles is the occurrence of steep vertical
gradients, with concentrations which can be orders higher near the
bottom than near the water surface. A representative illustration
of suspension concentration profile evolution by wave action over
coastal mudflats is provided by the data of Kemp and Wells (1987)
in Figure 27. Out of the four instantaneous (turbulence-mean),
vertical concentration profiles for suspended sediment, one
represents a pre-frontal (fairweather) condition, two during the
passage of a winter cold front and one post-frontal. The data were
obtained over a three day period at a site on the eastern margin
of the Louisiana chernier plain where the tidal range is less than
0.5 m. Wave height during front passage was on the order of 13 cm
and period 7 s. Of particular interest is the development of a
near-bed, high concentration suspension layer by the wind-generated
waves (profiles 2 and 3), which was previously absent (profile 1).
The post-frontal profile 4 further suggests that this layer may
have persisted following the front, conceivably due to the
typically low rate at which such a layer dewaters. The suspension
concentration in the upper water column was higher following the
front than that during the front, possibly due to sediment
advection from a neighboring area of higher turbidity.
Concentration profiles qualitatively similar to those shown
in Figure 27 have been reproduced in laboratory flume tests
involving wave action over soft muddy deposits (Ross and Mehta,
1989). Near the bottom, due to the entrainment of the mud/water
interface resulting from the effects of upward, shear-induced
diffusion strongly stabilized by the negative buoyancy of the high
concentration suspension, combined with hindered gravitational
settling, a significant lutocline develops. The height at which
the lutocline stabilizes is largely determined by a balance between
the rate of turbulent kinetic energy input and the buoyancy flux
determined by the sediment settling rate. In the water column above
the lutocline diffusion due to the wave field is characteristically
slow, so that the concentration there increases to modest levels
The formation of a high concentration fluidized layer of
sediment at the bottom is characteristic of all wave-influenced
environments having mud beds, including lakes (Wolanski et al.,
1989). In lakes such layers are episodically generated, but due to
the relatively low rates of dewatering of organics-rich sediment,
they are believed to be more common and persistent than thought
previously. As noted in Figures 4(a),(b), there is therefore an
obvious linkage between resuspension and trophic level in lakes
(see also Ostubo and Muraoka, 1986).
Gleason and Stone (1975) reported a concentration value of 102
mg/l at the water surface during a storm in the southern part of
Lake Okeechobee, Florida. In this lake, mud thickness does not
exceed about 80 cm (Figure 28(a)), and is typically less but
continuous in the muddy zone (Figure 28(b)), which is largely
contained in the deeper (> 3 m) part of the waterbody (Figure 29)
(Kirby et al., 1989). By assuming the entire water column of 4.6 m
depth at the site to have a vertically uniform concentration of
102 mg/l, they arrived at an erodible bed thickness of 2.3 mm (see
Figures 30(a),(b)), which however seems unrealistically small. On
the other hand, based on examinations of bottom cores from the
lake, Gleason and Stone concluded that a "fluid zone" comprising
of sediment deposit of thickness on the order of 7-20 cm probably
occurs near the bed in this lake. Since fluidized mud is easily
entrained by waves, it is instructive to determine the depth of
erosion by considering sediment erosion/deposition due to wave-
induced bottom stress in order to ascertain the significance of the
fluid zone in relation to turbidity generation and mud erosion
In Figure 31, simulated concentration profiles using a 1-D
vertical model similar to that used by Costa (1989) noted
previously are shown (Hwang, 1989). The main difference between the
two models is in the expression for the vertical mass diffusivity;
Costa used the buoyancy stabilized form of the well known Prandtl-
von Karman expression appropriate for open channel flows, while for
the simulation shown in Figure 31, Hwang used the corresponding
expression applicable to waves. The necessary sediment related
input (settling velocity; see e.g. Figure 8(c), and bed erosion
rate relationship) was obtained from laboratory tests using bottom
material from Lake Okeechobee. Bed density profile required for
simulation was derived through in situ vibracoring (see Figure 32,
which shows a radiograph of a core sample). The water depth used
for this simulation is 2.8 m, wave height 0.8 m and period 3.5 s.
The simulated evolution of the concentration profile in
Figure 31 starting from a clear water column representing pre-
storm calm shows a strong qualitative resemblance to prototype data
(e.g. Figure 27). The concentration C3 = 4.4 g/l corresponds to
that value above which settling of this particular sediment was
found to be hindered (see Figure 8(c)). The value C1 = 122 g/l
(corresponding to a bulk density of 1.065 g/cm3) is that above
which the material possessed measurable vane shear strength(see
Figure 33), thus effectively characterizing it as a bed. Thus the
range 4.4 to 122 g/l identifies for this sediment the lower and
upper concentrations and corresponding elevations, respectively,
of the fluid mud layer. At the end of 2.5 hr when the surface
concentration reached about 0.1 g/l, a typical value during storm
wave action in the lake, a 14 cm thick fluid mud layer was formed.
At the same time the bed eroded 1.7 cm.
As shown for illustration in Figure 34, bottom concentration
(density) profile relative to the mud surface in the lake confirmed
the presence of a persistent fluid mud layer of the same order
(11 cm) as that suggested by the simulation in Figure 31.
The scope of this lecture precludes additional demonstration
of practical applications of fine sediment transport principles,
which applications have by now become quite extensive in coastal
environmental work. For a brief review of the available
methodologies, reference may be made to Mehta et al. (1989b).
Suffice it to note that while considerable additional work remains
to be done in developing an understanding of the processes of
higher than present order of accuracy, the technology has come a
long way in the recent decades, and continues to advance with the
availability of better instrumentation for field and laboratory
A great many number of challenging engineering problems
involving fine sediments actually occur in developing countries
where, for instance, port development continues to remain a high
priority issue owing to obvious commercial imperatives. These
countries are important users and recipients of fine sediment
transport-related technology, which is being pursued through
research and development in a handful of industrially advanced
nations. In these latter countries, U.S.A. and Japan among them,
research and development in fine sediment transport in the coming
years will be guided by technical policy issues arising from three
main concerns, the foremost being water quality and contaminant
transport problems in coastal waters. This priority should also
lead to new insights into the nature and the dynamics of the mud-
water interface and exchange processes at this interface. Secondly,
older and more conventional issues associated with sedimentation
in ports and navigation channels will continue to demand attention
due to the large and rising dredging costs and the need to minimize
the number of shallow water dump sites for dredged material
disposal. Finally, the need for considerable improvement in the
prediction of the response of muddy coast shorelines and wetlands
to global climatic changes, in particular sea level rise, will
demand new thinking related to fine sediment process principles
associated with small transport rates over very long time scales.
There will be no shortage of challenging applications.
Let me conclude by once again acknowledging TOA Corporation,
whose sponsorship made this presentation possible.
Burt T.N. (1986). Field settling velocities of estuary muds. In:
Estuarine Cohesive Sediment Dynamics, A.J. Mehta ed.,Springer-
Verlag, New York, 126-150.
Costa R.C.F. (1989). Flow-fine sediment hysteresis in sediment-
stratified coastal waters. M.S. Thesis, University of Florida,
Dyer K.R. (1986). Coastal and Estuarine Sediment Dynamics. Wiley,
Dyer K.R. (1989). Sediment processes in estuaries: future research
requirements. Journal of Geophysical Research, 94(C10), 14327-
Futawari T., Kusuda T., Koga K., Araki H., Umita T. and Furumoto
K. (1988). Development of a new simulation method for suspended
sediment transport in a tidal river. Wat. Sci. Tech., 20(6/7), 103-
Gleason P.J. and Stone P.A. (1975). Prehistoric trophic level
status and possible cultural influences on the enrichment of Lake
Okeechobee. Report, South Florida Water management District, West
Grim R.E. (1968). Clay Mineralogy, McGraw-Hill, New York.
Halliwell A.R. and O'Dell M. (1969). Differences in silt patterns
across an estuary. The Dock and Harbor Authority, L(585), 125.
Hayter E.J. and Mehta A.J. (1986). Modelling cohesive sediment
transport in estuarial waters. Applied Mathematical Modelling, 10,
Hwang K.-N. (1989). Erodibility of fine sediment in wave-dominated
environment. M.S. Thesis, University of Florida, Gainesville.
Japan International Cooperation Agency (1988). A case study on
siltation countermeasures at Kumamoto port, Japan.
Jenkins S.A., Inman D.L. and Bailard J.A. (1980). Opening and
maintaining tidal lagoons and estuaries. Proc. Seventeenth Coastal
Engineering Conference, ASCE, New York, 1528-1547.
Kemp G.P. and Wells J.T. (1987). Observations of shallow water
waves over a fluid mud bottom: implications to sediment transport.
Proc. Coastal Sediments'87, ASCE, New York 367-377.
Kirby R. (1986). Suspended fine cohesive sediment in the Severn
estuary and Inner Bristol channel, U.K. Rept. ETSU-STP-4042, Dept.
of Atomic Energy, Harwell, U.K.
Kirby R.R., Hobbs C.H. and Mehta A.J. (1989). Fine sediment regime
of Lake Okeechobee, Florida. Rept. UFL/COEL-89/009, Coastal and
Oceanographic Engineering Dept., University of Florida,
Krone R.B. (1962). Flume studies of the transport of sediment in
estuarial processes. Final Rept., Hydraulic Engineering Laboratory
and Sanitary Engineering Research Laboratory, University of
Krone R.B. (1963). A study of theological properties of estuarial
sediments. Tech. Bull. No.7, Committee on Tidal Hydraulics, U.S.
Army Engineer Waterways Experiment Station, Vicksburg, Mississippi.
Kynch G.J. (1952). A theory of sedimentation. Transactions of the
Faraday Society, 48, 166-176.
Lick W. (1982). Entrainment, deposition and transport of fine-
grained sediment in lakes. Hydrobiologia, 91, 31-40.
Lijklema L., Gelencser P. and Szilagyi F. (1983). Sediments and
sediment-water interaction. In: Eutrophication of Shallow Lakes:
Modeling and Management, The Lake Balaton Case Study, L. Somlyody,
S. Herodek and J. Fisher eds.,International Institute for Applied
Systems Analysis, Laxenburg, Austria, 81-100.
Lin C.-P. and Mehta A.J. (1989). Turbidity-induced sedimentation
in closed-end channels. Journal of Coastal Research, 5(3), 391-
Maa P.-Y. (1986). Erosion of Soft Muds by Waves. Ph.D.
Dissertation, University of Florida, Gainesville.
Maa P.-Y. and Mehta A.J. (1985). Fine sedimentation in small
harbors. In: Flocculation, Sedimentation and Consolidation, B.M.
Moudgil and P. Somasundaran eds., Engineering Foundation, New York,
Maa P.-Y. and Mehta A.J. (1987). Mud erosion by waves: a laboratory
study. Continental Shelf Research, 7(11/12), 1269-1284.
Maa P.-Y. and Mehta A.J. (1989). Considerations on soft mud
response under waves. In: Estuarine Circulation, B.J. Nielsen, A.
Kuo and J. Brubaker eds., Humana Press, Clifton, New Jersey, 309-
McAnally W.H., Letter J.V., Stewart J.P., Thomas W.A. and Brogdon
N.J. (1984a). Columbia River hybrid modeling system. Journal of
Hydraulic Engineering, 110(3), 300-311.
McAnally, W.H., Thomas W.A., Letter J.V. and Stewart J.P. (1984b).
The Atchafalaya River delta, report 6, interim summary report of
growth predictions. Tech. Rept. HL-82-15, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, Mississippi.
Mehta A.J. (1973). Depositional Behavior of Cohesive Sediments.
Ph.D. Dissertation, University of Florida, Gainesville.
Mehta A.J. (1988). Laboratory studies on cohesive sediment erosion
and deposition. In: Physical Processes in Estuaries, J. Dronkers
and W. van Leussen eds., Springer-Verlag, New York, 427-445.
Mehta A.J. and Lott J.W. (1987). Sorting of fine sediment during
deposition. Proc. Coastal Sediments'87, ASCE, New York, 348-362.
Mehta A.J. (1989a). On estuarine cohesive sediment suspension
behavior. Journal of Geophysical Research, 94(C10), 14303-14314.
Mehta A.J. (1989b). Fine sediment stratification in coastal waters.
Proc. Third National conference on Dock & Harbour Engineering,
Karnataka Regional Engineering College, Surathkal, India, 487-492.
Mehta A.J., Hayter E.J., Parker W.R., Krone R.B. and Teeter A.M.
(1989a). Cohesive sediment transport I: Process Description.
Journal of Hydraulic Engineering, 115(8), 1076-1093.
Mehta A.J., McAnally W.H., Hayter E.J., Teeter A.M., Schoellhamer
D., Heltzel S.B. and Carey W.P. (1989b). Cohesive sediment
transport II: application. Journal of Hydraulic Engineering,
Mimura N. (1989). Recent Japanese studies on cohesive sediment
transport. Journal of Coastal Research, Special Issue No.5, 101-
Odd N.V.M. and Cooper A.J. (1989). A two-dimensional model of the
movement of fluid mud in a high energy turbid estuary. Journal of
Coastal Research, Special Issue No. 5, 185-193.
Oduyemi K.O.K. (1986). Turbulent transport of sediment in
estuaries. Ph.D. Dissertation, University of Birmingham,
Otsubo K. and Muraoka K. (1987). Field studies on physical
properties of sediment resuspension in Lake Kasumigaura. The
Japanese Journal of Limnology, 48 (Special Issue), 131-138.
Parchure T.M. and Mehta A.J. (1985). Erosion of soft cohesive
sediment deposits. Journal of Hydraulic Engineering, 111(10), 1308-
Parker W.R. and Kirby R. (1982). Time dependent properties of
cohesive sediment relevant to sedimentation management- A European
experience. In: Estuarine Comparisons, V.S. Kennedy ed., Academic
Press, New York, 573-590.
Parchure T.M. and Mehta A.J. (1985). Erosion of soft cohesive
sediment deposits. Journal of Hydraulic Engineering, 111(10), 1308-
Richardson J.F. and Jeronimo M.A. da S. (1979). Velocity-voidage
relations for sedimentation and fluidization. Chemical Engineering
Science, 34, 1419-1422.
Ross M.A. (1988). Vertical structure of estuarine fine sediment
suspensions. Ph.D. Dissertation, University of Florida,
Ross M.A. and Mehta A.J. (1989). On the mechanics of lutoclines
and fluid mud. Journal of Coastal Research, Special Issue No. 5,
Salomons W. and Mook W.G. (1977). Trace metal concentrations in
estuarine sediments: mobilization, mixing or precipitation. Journal
of Sea Research, 11(2), 119-129.
Sakakiyama T. and Bijker E.C. (1989). Mass transport velocity in
mud layer due to progressive waves. Journal of Waterway, Port,
Coastal and Ocean Engineering, 115(5), 614-633.
Scarlatos P.D. and Mehta A.J. (1990). Some observations on erosion
and entrainment of estuarine fluid muds. In: Physics of Shallow
Estuaries and Bays, R.T. Cheng ed., Springer-Verlag, New York (in
Sills G.C. and Elder D.McG. (1986). The transition from sediment
suspension to settling bed. In: Estuarine Cohesive Sediment
Dynamics, A.J. Mehta ed., Springer-Verlag, New York, 192-205.
Srinivas R. (1989). Response of fine sediment-water interface to
shear flow. M.S. Thesis, University of Florida, Gainesville.
Srivastava M. (1983). Sediment deposition in a coastal marina.
Rept. UFL/COEL/MP-83/1, Coastal and Oceanographic Engineering
Dept., University of Florida, Gainesville.
Tsuruya H., Murakami K. and Irie I. (1990). Supercomputer
simulations of mud transport by a multilayered nested grid model.
Paper to be presented at the Twenty-second International Conference
on Coastal Engineering, ASCE, Delft, The Netherlands.
Umita T., Kusuda T., Awaya Y., Onuma M. and Futawari T. (1984).
Wat. Sci. Tech., 17, 915-925.
van Olphen H. (1963). An Introduction to Clay Colloid Chemistry.
Wiley-Interscience, New York.
Wanless H.R. (1975). Sedimentation in canals. Report, Division of
Marine Geology and Geophysics, Rosenstiel School of Marine and
Atmospheric Science, University of Miami, Florida.
Williams P.R. and Williams D.A.J. (1989a). Rheometry for
concentrated cohesive suspensions. Journal of Coastal Research,
Special Issue No. 5, 151-164.
Williams D.A.J. and Williams P.R. (1989b). Rheology of concentrated
cohesive sediments. Journal of Coastal Research, Special Issue No.
Wolanski E., Asaeda T. and Imberger J. (1989). Mixing across a
lutocline. Limnology and Oceanography, 34(5), 931-938.
Plume (Negative Buoyancy)
<~ '"- Fluid Mud Layers
Response Time for Bed Formation Process for Fine
Sediment > Response for Sand.
Figure 1: Dumping of dredged fine material in coastal waters leads to sediment
spreading due to slow rate of settling and dewatering.
S' Individual (
Particle (1 prm)
Figure 2: Individual clay particles and agglomerated structure.
0 20 40 60 80 100
% < 16 MICRONS (particle size)
Figure 3: Variation of heavy metal concentration with sediment particle size (after
Salomons and Mook, 1977).
i 100 '
Anthropogenic Effects (A)
-------- I---- -----
Critical, Very Complex
Links that Require
Figure 4(a): Linkage between macroscale changes in coastal topography and mi-
croscale near-bed processes.
of Lake (A)
Figure 4(b): Replacing boxes (A) and (B) in Fig. 4(a) with these boxes defines the
problem of lake eutrophication.
Bottom Response is Tide Controlled in
Macrotidal Areas Episodic (Largely) in
Figure 5: Accumulation of "fluff" at channel bottom due to residual near-bed mud
Bottom Response is Entirely Episodic /
Figure 6: Schematic description of phosphorus loading in lakes (adapted from Lijk-
lema et al., 1983).
course medium fine course medium fine Icourse medium fine
100 60 40 20 10
SRange Cohesion Cohesion
Important Very Important
Figure 7: Size distributions of fine sediments and
Suspension -- Solution
their relationship to cohesion (after
k ;2 U 3 C;4;
Figure 8(a): Variation of settling velocity and settling flux with suspension concen-
tration (after Costa, 1989).
V San Francisco Mud
I MI I I
0 Maracaibo Mud
I I I I
2/ ----Highly Cohesive (n =4/3)
Figure 8(b): Variation of exponent n for flocculation settling with the degree of
10"' 1 10
CONCENTRATION, C (gL')
Figure 8(c): Settling velocity data from a laboratory settling column experiment using
sediment and fluid from Lake Okeechobee, Florida (after Hwang, 1989).
0 1 2 3 4 5 6
SHEAR STRAIN RATE (sec1)
Figure 9: Pseudoplastic shear stress-rate of strain relationships for a kaolinite bed at
different depths below bed surface (after Maa, 1986).
Figure 10: Schematic depiction of response of a mud bed to waves. Box with dashed
line (- -) represents the two-layered system, 7 = forcing function, 0 = components
of the two-layered system, 0 = transfer functions and O = response (after Maa and
Mehta, 1989). 1r
Bed Surface (cm)
Deforming Cohesive Bed
Stationary Cohesive Bed
Figure 11(a): Horizontal layering of vertical concentration
initions (after Mehta, 1989b).
profile and associated def-
u = Mean Current
us= Sediment Transport Velocity
Figure 11(b): Mass flux coupling between concentration and velocity profiles. Note
that u,/i can be as low as 0.5.
Figure 12: Transition from particulate behavior to soil mechanical behavior associated
with change in concentration and flow regime.
c u Mixed Layer Mobile
SU_ . Settling
Fluidization Bed Formnation
-- Mobile Fluid Mud Non-Newtonian
CStationary Fluid Mud J
Figure 13: Vertical transport processes characterizing the dynamics of the concentra-
tion field (after Mehta, 1989a).
6.00- C (g/I)
-2.50 -1.50 -0.50 0.50 1.50 2.50
Figure 14: Simulated concentration-velocity (squared) hysteresis loop (after Costa,
6 2.75 m
-1.0 U 1.25 m
-2.5 -1.5 -0.5 0.5 1.5 2.5
Figure 15: Measured concentration-velocity (squared) hysteresis loop (after Costa,
Figure 16: Reynolds stress versus velocity hysteresis (after Costa, 1989).
Figure 17: Location map of the Atchafalaya Bay area, Louisiana (after McAnally et
I I 11
A Decelerating Flow
: / /
I I -4--W I 1
Figure 18: Predicted delta extent in Atchafalaya Bay in the year 2030: (a) by ex-
trapolation method, and (b) by quasi-two-dimensional modeling (after McAnally et
Sei mn Le
Sediment Laden .' .
Water .. -. .. ..
; I "Inflow: :.Deposition -
Figure 19: Sediment influx in a dead-end canal by turbidity current.
Figure 20(a): Orange Waterway, a closed-end residential canal off Marco River,
2 ..................................................... ...Tu .! d .C u rren....n t ................................... .............................................. .
.............. .................................... .. ........ ................. c m ..
6 ................... ............... ................................................................... .5..-....cm .Y r .......................... 3.6.., .c M V ... 2 5 .n.3.5
I 7Sand and Muddy Sand
= Gray Muds
M Brown Muds
Figure 20(b): Longitudinal deposition rates and variation in Orange Waterway
(adapted from Wanless, 1975).
C Test No. Measured WSl/uA Sediment
SWES-8 0.0005 Kaolinite
Sp COEL-4,5,11,12 0.0015-0.0033 Kaolinite
S10.0 COEL -3 0.0145 Kaollnite -
Z U COEL-2 0.0149 Kaolinite
1.0 n ,H- /u= 0.0008
Z U 0.0060
z 0.01 I I I I I I
S0 20 40 60 80 100 120
0 DIMENSIONLESS DISTANCE FROM ENTRANCE,x/H
Figure 21: Dimensionless mean concentration below interface as a function of dimen-
sionless distance along dead-end canal (after Lin and Mehta, 1989).
I I I I I I 1 I I
S = 0.015 o 0
S Test No.
Sa 0 0 COEL 2,3,4,5
6A COEL 6,7,8
I I I I 1ES-13,
I I I I I .
Cedar Key Mud-
MEAN CONCENTRATION AT CHANNEL
ENTRANCE, C1 (mg 1)
Figure 22: Mean sediment flux as a function of depth-mean concentration at the
entrance (after Lin and Mehta, 1989).
|0.02 0.006 0.002
0 25 50 75 100 125 150
CONCENTRATION, Co (mg/I)
Figure 23: Histogram showing frequency of occurrence of concentration outside a
basin (after Maa and Mehta, 1985).
Figure 24: Camachee Cove Yacht Harbor basin on Tolomato River near the Atlantic
coast of Florida.
CONCENTRATION, CoIN RIVER (mg/I)
Figure 25: Variation of shoaling or sedimentation rate, SR, with outside concentration
Co (after Maa and Mehta, 1985).
I I I
2-D Model Simulation
Mean SR = 0.146 m/yr
Figure 26: Predicted sedimentation contours in Camachee Cove Yacht Harbor basin
(after Hayter and Mehta, 1986).
Figure 27: Vertical suspended sediment concentration profiles obtained before, during
and after the passage of a winter cold front at a coastal site in Louisiana. Times are
relative to time of measurement of the pre-frontal profile 1 (adapted from Kemp and
Figure 28(a): Mud thickness contour map of Lake Okeechobee, Florida (after Kirby
et al., 1989).
f Acoustic Profile
Figure 28(b): A portion of a survey line demonstrating a relatively clean mud layer
over a harder substrate. The sub-bottom reflector depicts a small paleochannel show-
ing signs of some internal compaction (after Kirby et al., 1989).
Most Mud is
Bathymetry of Lake Okeechobee, Florida, with region having depths >
A Simplified More Realistic
Figure 30: a) Relationship between uniform suspension concentration, C,, in water
column of depth hi, and the corresponding thickness, h2, of bed of concentration
Cb; b) High concentration suspension layer between low concentration and bed (after
SEDIMENT CONCENTRATION (g/I)
Figure 31: Simulated lutocline evolution under wave action (after Hwang, 1989).
60% Clay, 40% Organic Matter
Hard "Beach Rock"
Lamination Suggests a
Figure 32: X-radiograph of core
Kirby et al., 1989).
OK11 VC from Lake Okeechobee, Florida (after
1.065 g cni3
BULK DENSITY (g cm3)
Figure 33: Mud vane shear strength variation with density (after Hwang, 1989).
100 200 300 400 500
CONCENTRATION, C (g/I)
Figure 34: Bottom concentration profile showing 11 cm thick fluid mud layer (after
- ---- I I I
cQ-3 4.4 g/I
7 i/ =22gi,-
110 = 1122 g/I