-------
•A-FRAME
*iKy&«lSS^^
CUTTER HEAD
a. Cutterhead pipeline dredge
jte^^
b. Bucket dredge
c. Self-propelled hopper dredge
Figure 1. Hydraulic cutterhead (a), bucket (b), and hopper (c) dredges (from Palermo and
Pankow 1988)
Chapter 2 Contaminant Losses During Dredging
11
-------
packed deposits and cut through soft rock, it can excavate a wide range of
materials including clay, silt, sand, and gravel.
The cutterhead dredge is suitable for maintaining harbors, canals, and
outlet channels where wave heights are not excessive, allowing it to work
effectively in all types of alluvial sediments and compacted deposits. A cut-
terhead dredge is typically equipped with two stern spuds that alternately serve
as a pivot swinging the cutterhead from side to side during operation. Resus-
pension of sediments during cutterhead excavation is strongly dependent on
operational parameters such as thickness of cut, rate of swing, and cutter
rotation rate. Proper balance of operational parameters can result in sus-
pended sediment concentrations as low as 10 mg/t in the vicinity of the cut-
terhead (Hayes, Raymond, and McLellan 1984).
Dustpan dredge
The dustpan dredge is a hydraulic suction dredge that uses high pressure
water jets to loosen sediment for capture by suction. Dustpan dredges are
used primarily for dredging sandy sediments on inland rivers. Dustpan
dredges generate suspended solids plumes similar to cutterhead dredges.
Plume suspended solids averaged 3.8 times background concentrations during
removal of kepone-contaminated sediments from the James River, Virginia
(McLellan et al. 1989).
Matchbox suction dredge
A matchbox dredge is a suction dredge that eliminates the cutterhead and
water jets used in other hydraulic pipeline dredges. The dredge was originally
designed to remove contaminated sediments in Rotterdam Harbor, The Neth-
erlands (Hayes, McLellan, and Truitt 1988). The dredge head is designed to
remove sediments close to in situ density and minimize resuspension. The
absence of mechanical mixing associated with a cutterhead or water jet should
reduce the sediment resuspension rates. However, the limited field data avail-
able indicate that paniculate release rates for the matchbox and cutterhead
dredges are about the same (Hayes, McLellan, and Truitt 1988; McLellan
et al. 1989). Sediment resuspension with both types of dredges is highly
dependent on operator skill and experience. In the studies conducted by
Hayes, McLellan, and Truitt (1988), operator inexperience with the matchbox
dredge contributed to poor control of matchbox position and frequent clogging
of the suction line.
Hopper dredge
Hopper dredges (Figure 1) are usually self-propelled vessels equipped with
dredge pumps for removing sediments and large hoppers for storing dredged
material during transportation. Sediment is raised by dredge pumps through
1 2
Chapter 2 Contaminant Losses During Dredging
-------
drag arms connected to drag heads and discharged to hoppers built in the
vessel. Dredging past hopper overflow is (i.e., allowing water and fine-grain
sediment particles to flow over and out of hoppers) sometimes practiced to
improve dredging economics by trapping coarse-grain material in the hoppers
and releasing fine-grain material in the overflow. Since contaminants tend to
be associated with fine-grain material, overflow is not a recommended dredg-
ing alternative for remediation. However, hopper dredging (without over-
flow) is an alternative that should be considered for sites requiring good
maneuverability and minimum interference with navigation.
Horizontal auger dredge
A horizontal auger (HA) dredge is a cutterhead suction dredge with hori-
zontal cutter knives and a spiral auger that cuts the material and moves it to
the suction. The dredge is designed for the removal of small amounts of
sediment (50 to 120 yd3/hr (Averett et al. 1990)). Nawrocki (1974) reported
resuspended sediment concentrations two to four times background within 4 m
(12 ft) of the auger between the surface and the bottom. In a pilot study in
New Bedford Harbor, the HA-type dredge experienced problems with posi-
tioning, anchoring, and effectiveness of the mudshield. Sediment resuspension
at the dredgehead was substantially higher than for either cutterhead or match-
box dredges.
Cleanup dredge
Sato (1976a,b) describes an instrumented, covered auger dredge that is
designed to clean up highly contaminated sediments. The instrumentation
includes a sonar to determine the bottom elevation and an underwater televi-
sion camera for monitoring of dredging operations. Resuspended sediment
concentrations observed during tests of this system were essentially indistin-
guishable from background levels (Herbich and Brahme 1991).
Pneuma pump
The Pneuma pump uses compressed air and hydrostatic pressure rather
than centrifugal motion to move dredged material through a pipeline. During
the dredging process, the pump is submerged and sediment and water are
forced into one of three cylinders by opening the cylinder to atmospheric air.
The pump must be used at depths in excess of approximately 4 m (12 ft) to
provide sufficient hydrostatic pressure for effective filling. After filling,
compressed air is supplied, forcing the water and sediment through an outlet
valve. Richardson et al. (1982) conducted field tests on a Pneuma pump and
observed low turbidity levels in the vicinity of the pump. It was not possible
to dredge sand, and the hydraulic efficiency of the dredge was consistently
below 20 percent. Barnard (1978) reported suspended solids concentrations
an order of magnitude above background within 1 to 2 m of a Pneuma pump.
Chapter 2 Contaminant Losses During Dredging
-------
Oozer pump
The Oozer pump operates in a manner similar to the Pneuma pump, but
vacuum is applied during the filling stage to achieve more rapid rilling,
increase solids concentrations, and allow operation in more shallow waters.
In the 11-year period from 1974 to 1984, approximately one million cubic
meters of contaminated sediment were removed by the Oozer dredge
(Ikalainen 1987). The Japanese Dredging and Reclamation Engineering Asso-
ciation conducted a field test of the Oozer dredge in Osaka Bay, Japan. The
Oozer dredge removed organically contaminated fine-grained sediment in
16 m of water. Results indicated that the primary source of sediment resus-
pension around the Oozer dredge is the swing speed. Suspended solids con-
centrations were monitored at locations 50, 100, 200, and 300 m in front of
the Oozer dredgehead; three sample stations were radially located at these
distances. The maximum concentration observed at the three stations was
14 mg/f (Zappi and Hayes 1991).
Bucket dredge
A bucket dredge is a mechanical device that utilizes a bucket to excavate
sediment (Figure 1). Unlike hydraulic dredges that typically remove four
times as much water as in situ sediment, the bucket dredge can remove mate-
rial at close to in situ densities. It is used near surface and submerged struc-
tures due to the greater degree of control allowed during dredging. It can also
be used to dredge at greater depths than many hydraulic dredges. Most of the
contaminant losses during bucket dredging occur during the impact, penetra-
tion, and removal of the bucket from the sediment (Hayes, McLellan, and
Truitt 1988). Significant losses also occur during hoisting through the water
column and after the bucket breaks the surface due to drainage from the
bucket. Palermo, Homziak, and Teeter (1990) estimated that 20 to 30 percent
of the sediment excavated from a clay and silt bed was spilled from a clam-
shell bucket before reaching the disposal scow. These losses can be mini-
mized through operational controls and the use of enclosed buckets (Barnard
1978). Operational controls include smooth hoisting of the bucket and use of
a hoisting speed less than 2 m/sec (McLellan et al. 1989).
Other mechanical dredges include the backhoe, bucket ladder and wheel,
dipper, and dragline dredges. All of these dredges are expected to increase
the amount of resuspended sediment over the bucket dredge (Averett et al.
1990). The backhoe, bucket ladder and wheel, and dragline dredges are not
appropriate for remediation dredging, which is the focus of this chapter.
Paniculate Contaminant Releases During Dredging
The discussion below will outline procedures for estimating contaminant
losses from those dredges for which predictive techniques have been
14
Chapter 2 Contaminant Losses During Dredging
-------
proposed, specifically cutterhead and bucket dredges. These predictive tech-
niques are based on limited studies and, therefore, not fully developed nor
verified. Because additional studies involving a wide range of dredging condi-
tions are needed, the proposed predictive techniques for losses during dredg-
ing should be regarded as unproven techniques requiring additional research
and development.
An alternative approach is use of the sediment resuspension information
compiled in Table 2 (from Nakai 1978 as cited by Herbich and Brahme 1991).
This information provides rough guidelines for estimating resuspension rates
by cutterhead and bucket type dredges. The guidance provides insufficient
information, however, to indicate the effect of operational controls or the
influence of different types of sediments.
Table 2
TGU's1 for Different Dredges and Dredging Projects (Nakai 1978}2
Type of
Dredge
Pump
Trailing suction
Grab
Bucket
Installed
Power or
Bucket
Volume
4,000 hp
2,500 hp
2,000 hp
2,400 hp
x2
1,800 hp
8 cu m
4 cu m
3 cu m
Dredged Material
(d <74|/, %)3
99.0
98.5
99.0
31.8
69.2
74.5
94.4
3.0
2.5
8.0
92.0
88.1
83.2
58.0
54.8
45.0
62.0
87.5
10.2
27.2
d <5fj. %
40.0
36.0
47.5
11.4
35.4
50.5
34.5
3.0
1.5
2.0
20.7
19.4
33.4
34.6
41.2
3.5
5.5
6.0
1.5
12.5
Classification4
Silty clay
Silty clay
Clay
Sandy loam
Clay
Sandy loam
Silty clay
Sand
Sand
Sand
Silty clay loam
Silty loam
Silt
Silty clay
Clay
Silty loam
Silty loam
Silty loam
Sand
Sandy loam
TGU
kg/m3
5.3
22.5
36.4
1.4
45.2
12.1
9.9
0.2
3.0
0.1
7.1
12.1
25.2
89.0
84.2
15.8
11.9
17.1
17.6
55.8
1 TGU = kilograms of suspended sediment per cubic meter material dredged.
2 Nakai (1978) as cited by Herbich and Brahme (1991).
3 d = diameter of soil particles.
4 Classification is according to the triangular soil classification system.
Chapter 2 Contaminant Losses During Dredging
15
-------
General considerations
Resuspension of particulates is a function of dredge type and operation and
sediment properties. The effects of operational factors on resuspension for
selected dredges will be reviewed here. Sediment properties are a site-specific
concern that cannot be definitively quantified without reference to a specific
dredging project. In general, finer, less cohesive sediments have the greatest
potential for resuspension.
Contaminants associated with resuspended particulates are primarily metals
and other elemental species and organic contaminants. Elemental species of
concern may be in geochemical phases with slow release properties or in
geochemical phases that readily accept and release elemental species. Organic
contaminants are usually bound in the organic fraction of the sediment through
reversible sorption reactions. Contaminant species may also be dissolved in
the pore water adjacent to the sediment particles; but for most contaminants,
the dissolved fraction is much smaller than the particulate fraction.
The mass release of a contaminant during dredging is defined by
m=frpsADCs (1)
where
m = contaminant mass released, g
fr = fraction of sediment resuspended during dredging, dimensionless
ps = in situ bulk density of the sediment, g/cm3
A = dredging area available for mass transfer, cm2
D = dredging depth, cm
Cs = contaminant concentration in sediment (dry wt), g/g
Equation 1 is useful as a definition, but it is not as a predictive equation
because the fraction of sediment resuspended is difficult to estimate and mass
release is more conveniently expressed on a rate basis. To obtain the rate of
mass release, the dredging area, A, is replaced with Ad, the area of dredging
per unit time (square centimeters per second) and m becomes RD, the mass of
contaminant released per unit time (grams per second). Alternatively, if an
average water column resuspended solid concentration is known over some
volume, the rate of contaminant resuspension, RD, is given by
R = C Q C (2)
16
Chapter 2 Contaminant Losses During Dredging
-------
where
RD = rate of contaminant release, g/sec
Cp = suspended solids concentration averaged over a
characteristic volume at point of dredging, g/cm3
Qd = volumetric flow of water through averaging volume, cnrVsec
Figure 2 shows a definition sketch for Equation 2. It should be noted that the
bulk sediment contaminant concentration is generally reported as mass of
contaminant per mass of dry sediment and implicitly assumes that all the
contaminant mass resides on the solid phase. The contaminant release rate
defined in Equation 2 is based on the total contaminant concentration initially
in the in situ sediment and, therefore, includes both particulate and dissolved
contaminant fractions.
AVERAGING
VOLUME
Cp = AVERAGE
RESUSPENDED
SEDIMENT
CONCENTRATION
Qd > VOLUMETRIC
FLUSHING RATE
'^^^^^$^^^^
BOTTOM
SEDIMENT
Figure 2. Definition sketch for contaminant release at point of dredging
Estimation of the total contaminant release or the release rate per unit time
by resuspension of the sediment is thus reduced to estimation of the fraction of
particles that are resuspended. The rate of sediment resuspension is discussed
for cutterhead hydraulic and bucket dredges in the sections that follow. The
dissolved fraction of the total contaminant loss will be discussed in a later
section.
Chapter 2 Contaminant Losses During Dredging
17
-------
Cutterhead dredges
Cutterhead dredges loosen the bottom sediment by the mechanical action of
the multiblade rotating cutterhead. The sediment dislodged in this manner is
drawn via hydraulic suction into the suction pipe and transported to the dis-
posal site by pipeline. Paniculate contaminant release occurs during this
process when the hydraulic suction is unable to completely entrain all of the
dislodged sediment.
A controlled study of particulate releases by a cutterhead dredge was con-
ducted by Hayes, McLellan, and Truitt (1988) at Calumet Harbor, Illinois.
Key operational parameters that affected sediment resuspension rates were the
rotation rate of the cutterhead and swing speed of the cutterhead ladder on
which the cutterhead is supported. Overcutting, when the sense of rotation of
the cutterhead and the ladder are the same (Figure 3), resulted in higher rates
of sediment resuspension. During overcutting, the shear of the cutterhead
relative to the water is greatest when the cutterhead is at the top of its rota-
tion, resulting in more resuspension of dredged material on the cutterhead.
As shown in Figure 3, undercutting (which occurs when the sense of cutter-
head at the top of its rotation and ladder swing differ) reduces sediment resus-
pension. This explanation of cutterhead resuspension is consistent with the
concept that the tangential velocity of the cutterhead relative to the essentially
motionless water is the primary factor in sediment resuspension.
LEFT SWING
RIGHT SWING
ENTRAINED WATER
SEABED
SEDIMENT PARTICLE
CUTTER REVOLVING DIRECTION
CUTTER BLADE
DREDGED MATERIAL
REMAINING
SUCTION MOUTH
DREDGING SEDIMENT THICKNESS
UNDERCUTTING
OVERCUTTING
Figure 3. Cutting operation of a cutterhead dredge (front view)
During overcutting, the effective blade velocity is the sum of the tangential
velocity of the cutterhead blades about their axis of rotation and the swing
velocity of the dredge, that is, the velocity of the ladder with respect to the
dredge. During undercutting, the effective blade velocity is the difference
between the two velocities. Additional factors controlling cutterhead resus-
pension include the degree of head burial in the sediment and the characteris-
tic velocity of the cutterhead intake. Increases in the intake velocity reduce
the fraction of the particles that are resuspended by the cutterhead but not
removed by the hydraulic suction.
18
Chapter 2 Contaminant Losses During Dredging
-------
Fully buried cutterheads reduce the exposure of the loosened sediments to
the overlying water and therefore increase the fraction that are removed by the
hydraulic suction.
Hayes (1986) correlated resuspended sediment concentrations to powers of
the dimensionless ratios of swing velocity-to-characteristic intake velocity and
cutterhead tangential velocity-to-characteristic intake velocity. Collins (1989)
restated this correlation in the manner given in Equation 3.
c_
P
= FFFD
w
V.
V:
V.
(3)
where
Cp = suspended solids concentration averaged over a characteristic volume
at point of dredging, g/cm3
pw = density of water, g/cm3
FF = coefficient for all factors other than degree of burial, dimensionless
FD = cutterhead resuspension rate factor accounting for degree of burial,
dimensionless
Vs = swing velocity of dredge ladder, cm/sec
Vj = characteristic velocity of cutterhead intake, cm/sec
a = empirical swing velocity significance factor, dimensionless
Vc = effective blade velocity, cm/sec
b = empirical tangential velocity significance factor, dimensionless
FD = 1 for full cut (fully buried cutterhead) dredging and would be greater
than 1 for partially buried dredging. FF is a site-specific factor that accounts
for sediment and dredge operational variations. Based on 12 data sets with a
fully buried cutterhead at Calumet Harbor, Illinois, Hayes (1986) found that
a = 2.85, b = 1.02, and FF = 0.089 with a correlation coefficient of 0.72.
Collins (1989) extended this correlation for Calumet Harbor to other sites
and dredging conditions. Using data collected during cutterhead dredging
operations at Calumet Harbor, Illinois, Savannah River, South Carolina, and
James River, Virginia, Collins (1989) developed the following predictive
equations for the factors FF and FD.
19
Chapter 2 Contaminant Losses During Dredging
-------
log10FF = lo- -2.05
FD = 1 + 1.9(£>F-1)2 + 0.41(Z)F-1)7 (5)
where
Dch = diameter of cutterhead, cm
d = effective diameter of sediment grains, cm
DF = fractional depth of cut as a function of cutterhead diameter,
dimensionless
Equation 5 is a modification of Equation 30 in the report of Collins (1989) to
reflect the fact that FD - 1 for DF = 1 . Using all of the data for the three
sites, the correlation coefficient for the model is 0.556.
Use of the correlation given by Equation 4 is hindered by its sensitivity to
the ratio of the cutterhead diameter to the sediment grain-size diameter.
Changes in the average grain size of less than a factor of two can result in a
change in the factor, FF, by more than an order of magnitude. Extreme cau-
tion is warranted in the use of Equation 4. If sediment resuspension rates
estimated using Equation 4 differ by more than a factor of 10 from the
approximate estimates of Nakai (1978) (Table 2), the approximate estimates
given in Table 2 should be used.
Equations 3 through 5 provide an estimate of the resuspended sediment
concentration, Cp, a variable in Equation 2. In addition to the resuspended
sediment concentration, the volumetric flow of water through the characteristic
volume over which the resuspended solids concentration is averaged is
required. Collins (1989) approached obtaining the needed volumetric flow by
defining the characteristic volume for averaging as equal to the tangential
velocity of the cutting head relative to fixed coordinates (Vt) times the cross-
sectional area to which this velocity applies (Figure 2). Taking the height of .
the cutting head as Hch and its length as Lch, the cross-sectional area is aHch
@Lch where a and /3 account for fact that the sweep volume is typically larger
than the cutterhead. Collins' (1989) estimates of this volume are equivalent to
the values, a = 1.75 and /3 = 1.25. Additional field tests over a wide range
of dredging conditions will be needed before the general applicability of the a
and /3 values listed above can be fully evaluated.
Using the approach suggested by Collins (1989) for obtaining the volumet-
ric flow, the contaminant mass release rate equation (Equation 2) can be writ-
ten as
20
Chapter 2 Contaminant Losses During Dredging
-------
= CP
Despite its weaknesses, Equation 6 is the only equation presently available for
predicting contaminant release during cutterhead hydraulic dredging.
An alternative approach is the use of the sediment resuspension rates
observed by Nakai (1978) and summarized in Table 2. The tabulated resus-
pension rates could be used to indicate an approximate prediction and then
Equations 3 and 5 used to indicate the influence of cutterhead speed, swing
speed, and fractional depth of cut. This method is suggested under conditions
when Equation 4 indicates extreme sensitivity to sediment grain size.
Bucket dredges
Different types of bucket dredges can fulfill various types of dredging
requirements. Typical buckets include the clamshell, orange-peel, and drag-
line types. This discussion will focus on the clamshell type of bucket dredge.
Sediment is resuspended during bucket dredging operations by impact, pene-
tration, and withdrawal of the bucket and during hoisting of the bucket.
Bucket dredges usually excavate a heaped bucket of material, but, during
hoisting, a portion of the load washes away. Once the bucket clears the water
surface, additional losses may occur through rapid drainage of water and
slumping of the material heaped above the rim. Loss of material is also influ-
enced by the fit and condition of the bucket, the hoisting speed, and the prop-
erties of the sediment.
A special type of bucket, the enclosed clamshell bucket, has been devel-
oped to minimize loss of dredged material. The edges seal when the enclosed
clamshell bucket is closed, and the top is covered. A comparison of conven-
tional clamshell and enclosed clamshell bucket dredging operations indicated
that the enclosed clamshell generates 30 to 70 percent less turbidity in the
water column than typical buckets (Barnard 1978).
The key parameters affecting total resuspension rate are bucket size, cycle
time, and type of bucket. The cycle time, or the time required to drop, fill,
and withdraw the bucket, is a function of the rate of each of the individual
steps (impact, penetration, withdrawal, and hoisting). The speed at which
these steps are accomplished significantly influences sediment resuspension
rates.
A dimensionless parameter that scales with the bucket volume is defined by
Collins (1989) as
B . J-pvy" , £ (7,
21
Chapter 2 Contaminant Losses During Dredging
-------
where
B = Collins bucket parameter, dimensionless
hb = water depth, cm
Vcb = volume of clamshell bucket, cm3
Lbc - characteristic length of clamshell bucket, cm
The term in brackets in Equation 7 is the characteristic size of the clamshell
bucket recognizing that the bucket is approximately square on two sides and
triangular on the third.
Collins (1989) defines a dimensionless cycle time Tc as
Tc = ^ (8)
"b
where
Tc = dimensionless cycle time
v3 = settling velocity of a representative particle, cm/sec
Tcb = bucket cycle time, sec
hb/v3 is the time required for a representative sediment particle to fall over the
entire depth of the water column.
Unfortunately, insufficient data exist to relate resuspension or contaminant
release rates with both B and Tc. Collins (1989) used the ratio of Tc to B to
define a new dimensionless variable as the ratio of the bucket cycle time to the
time required for particles to settle the bucket distance.
Il = ¥* (9)
A regression analysis of the resuspended sediment concentrations for experi-
ments at St. John River, Black Rock Harbor, and Calumet River (Collins
1989) suggested the correlation
Cp - 0.0023 Pw *
lc
(10)
22
Chapter 2 Contaminant Losses During Dredging
-------
where C_ is the resuspended sediment concentration, gm/cm3.
Collins (1989) reports that the logarithmic equivalent of Equation 10 has a
correlation coefficient of about 0.98.
Estimation of the release rate requires that the concentration estimated from
Equation 10 be multiplied by the exchange rate of the volume swept by the
bucket. The volumetric sweep rate of the bucket should be proportional to the
square of the characteristic clamshell bucket length times the effective velocity
of the bucket. As with the cutterhead dredge, the area swept by the bucket
during insertion and withdrawal exceeds the bucket area. Bohlen (1978)
suggests that the sweep area is approximately two to three times the area of
the bucket. The effective velocity of the bucket is approximately h^r^. If
the concentration predicted by Equation 10 applies throughout the sweep area
and dredging cycle, then the particle resuspension rate is given by
\4WL- / lit. r* I (•!_
*,6 = 7 -^--^C = 7 Pw (0.0023) (Lftc)2_L
Tc* rc6
_B
7,
where 7 is the Bohlen sweep area correction factor (2 to 4), dimensionless.
The only equation presently available for predicting the solids resuspension
rate during bucket dredging is Equation 11. Contaminant release rate is given
by modifying Equation 11 to include the concentration of contaminant in the
sediment, Cs, as shown in Equation 12 below.
5 c (12)
T
s
As indicated previously, an enclosed clamshell dredge should reduce the con-
taminant release rate predicted by Equation 12 by 30 to 70 percent.
An alternative approach is the use of the sediment resuspension rates
observed by Nakai (1978) (Table 2). The high correlation coefficient of
Equation 10 suggests, however, that the approach of Equations 7-12 is the
best estimate available.
Dissolved Contaminant Releases During Dredging
Resuspension of sediment solids during dredging can also impact water
quality through the release of contaminants in dissolved form. Before resus-
pension, contaminant distribution between sediment solids and sediment pore
water is probably at equilibrium. Dredging exposes sediments to major shifts
in liquids/solids ratio and oxidation-reduction potential (redox). Because the
23
Chapter 2 Contaminant Losses During Dredging
-------
sediment solids are removed from the equilibrium conditions previously exist-
ing, there is a potential for change in the distribution of contaminant between
solid and aqueous phases. Initially upon resuspension, the bulk of the contam-
inants are sorbed to paniculate matter. As the resuspended particulate con-
centration is diluted by mixing with dredging site water, release of sorbed
contaminants to adjacent waters results in a continuous increase in the fraction
of contaminants that are dissolved.
It should be noted that the total release of contaminants at the point of
dredging is estimated by the equations of the previous section. The dissolved
release calculated by the methods of this section largely occurs after the mix-
ing and dilution of the resuspended sediments with the ambient waters. The
fraction of the contaminant associated with the particulate phase continues to
change as dilution reduces the particle concentration.
In this section, equilibrium partitioning is discussed as a predictive tech-
nique for dissolved organic contaminants. Because equilibrium partitioning of
organic contaminants is discussed in detail in Contaminant Losses During
Pretreatment in the section on leachate quality, details of equilibrium partition-
ing theory are not presented in this section. A pseudo-equilibrium partitioning
approach for estimating dissolved metals concentrations is discussed in Con-
taminant Losses During Pretreatment, but this approach is not recommended
for application to release of dissolved metals during dredging because the
rapid and pronounced change in redox and the complicated environmental
chemistry of metals make equilibrium approaches highly unreliable and
uncertain.
The most accurate predictive indicator of dissolved contaminant release
during dredging would be a fully researched and developed laboratory test that
reproduces the mixing and dilution processes that are observed in the water
column after resuspension of contaminated sediments. Such a test would
indicate sediment-specific effects on desorption rate and contaminant tendency
to desorb. The test would be especially important for elemental species, such
as heavy metals, that undergo complex reactions that are not easily predicted
by mathematical models. The test would also be important for strongly
sorbed hydrophobic organic species that may desorb slowly due to mass trans-
fer resistances.
DiGiano, Miller, and Yoon (1995) proposed an adaptation of the standard
elutriate test, a dredging elutriate test (DRET), for the purpose of predicting
dissolved contaminant releases. The DRET is preliminary (only one sediment
tested) and requires further development before a test of this type can be
adopted for routine application. The standard elutriate test (SET) was devel-
oped during the DMRP to predict contaminant release during open-water
disposal operations (Jones and Lee 1978). In the SET, water and sediment
are mixed in a proportion of 4:1, mixed for 30 min and allowed to settle for
1 hr. The modifications suggested by DiGiano, Miller, and Yoon (1995) were
designed to achieve a more realistic solids/water ratio (0.5 to 10 g/l) consis-
tent with conditions for resuspended sediment due to dredging. DiGiano,
24
Chapter 2 Contaminant Losses During Dredging
-------
Miller, and Yoon (1993) employed an aerated mixing time of 1 to 6 hr and a
settling time of 1 hr (0.5 to 24 hr were also investigated).
The DRET was evaluated by comparison to field dredging studies con-
ducted in New Bedford Harbor, Massachusetts. The DRET was found to be a
reasonable indicator of the soluble and total (soluble plus unsettled paniculate)
polychlorinated biphenyl (PCB) concentrations released during cutterhead or
matchbox suction dredging but underpredicted PCB concentrations when a
horizontal auger dredge head was used. Additional testing of DRET at a
number of sites is needed before the general applicability of the test can be
evaluated. The New Bedford Harbor studies involved a highly contaminated
sediment at an estuarine location. Extrapolation of the New Bedford Harbor
results to freshwater sites with one to two orders of magnitude lower contami-
nation levels is not technically defensible at this time.
In the absence of specific information to the contrary, it, therefore, seems
appropriate to use equilibrium partitioning to establish an upper bound on
dissolved organic concentrations at the point of dredging. However, equilib-
rium partitioning is usually a very conservative assumption. DiGiano, Miller,
and Yoon (1990) found that an equilibrium partitioning model did a good job
of predicting the soluble PCB concentrations. At low contaminant concentra-
tions, equilibrium partitioning between sediment and water can usually be
represented by a linear isotherm, that is, Csorb — KdCw, where Kd is a distri-
bution coefficient assumed independent of concentration. Here, Cw is the
water phase concentration and Csorb is the concentration of the contaminant
sorbed to the solid phase. The sorbed concentration in the solid phase is
usually assumed to be approximately equal to the bulk sediment contaminant
concentration Cs, so that, Csorb » Cs.
Using local equilibrium partitioning, the dissolved concentration is given
by
C C
C = * p (13)
where
Cw = aqueous phase contaminant concentration, mg/£
Cs = bulk contaminant concentration in sediment, mg/kg
Cp = suspended solids concentration averaged over a characteristic volume
at point of dredging, kg/ 1
Kd = contaminant-specific equilibrium distribution coefficient, t /kg
The distribution coefficient in Equation 13 can be determined in batch equilib-
rium tests or estimated using empirical relationships from the literature.
25
Chapter 2 Contaminant Losses During Dredging
-------
Procedures for measuring or estimating the distribution coefficient are
described in Appendix B.
The release rate for dissolved contaminants is the product of the dissolved
contaminant concentration averaged over the volume swept by the dredge and
the volumetric flow through the averaging volume. The dissolved contami-
nant release rate for a cutterhead dredge is thus given by
R^ch = Cw V, aHch $Lch (14)
Similarly, the dissolved contaminant release rate for a clamshell bucket dredge
is given by
*u = ypw (Lbf— cw (is)
rcb
Several limitations apply to Equations 14 and 15. First, there are little
field data for verification of these equations. Second, Equations 14 and 15 are
not applicable to estimation of dissolved metals releases. In addition, the
linear partitioning used in Equations 14 and 15 assumes dissolved phase con-
centrations much lower than the water solubility limit. Deviations from linear
partitioning might be expected when dissolved phase concentrations approach
50 percent of the solubility limit.
Further, the total contaminant release for cutterhead hydraulic and bucket
dredges is provided by Equations 6 and 12, respectively. Although dissolved
losses at the point of dredging represent a small fraction of the total loss for
strongly sorbing chemicals, some estimation of dissolved losses, such as pro-
vided in Equations 14 and 15, may be needed for transport models used to
assess impacts and risks and to compare the no-action alternative to dredging
and treatment/disposal alternatives. Finally, Equations 14 and 15 predict
dissolved concentrations at the point of dredging, not downstream dissolved
concentrations.
Although hydrophobic organic species often partition in the simple manner
discussed previously, the release of metals is much more complex. During
the development of the standard elutriate test, there was little correlation
observed between sediment bulk metal concentration and the dissolved metal
concentration at disposal sites or in the standard elutriate. In most cases,
dissolved metal concentrations in site water prior to and during disposal opera-
tions were about the same (Jones and Lee 1978). In some cases, dissolved
metal concentrations were higher in site water prior to disposal operation than
after disposal operations (Jones and Lee 1978). These results can often be
explained in terms of the aqueous environmental chemistry of iron. Many
sediments contain a large reservoir of reactive ferrous iron that readily reacts
with oxygen in site water to form amorphorous iron oxyhydroxides. Iron
oxyhydroxides tend to floe and scavenge metals. Thus, an adaptation of the
26
Chapter 2 Contaminant Losses During Dredging
-------
SET such as DRET is probably required to get reliable estimates of soluble
metal releases during dredging.
Closure on Losses During Dredging
It is clear from the previous discussion of losses during dredging that a
number of dredging equipment factors and interactions between sediment and
water are likely to be important in predicting contaminant losses. Prediction,
however, requires simplifying assumptions about the relative importance of
these factors and interactions, followed by major extrapolations about the
complex and transient conditions of the field environment. Field measure-
ments of resuspension and desorption at the point of dredging supported by
data on operational factors and ambient conditions are, therefore, essential to
better understanding of contaminant release rates at the point of dredging.
The number of such studies is rather limited. They are complex and expen-
sive, involving major investments in equipment (dredges) and chemical analy-
ses. It is important, therefore, that future studies be designed to provide the
maximum amount of information on relevant factors and interactions.
The predictive equations presented in this section may at first glance seem
straightforward and easy to apply. For many of the variables in the equations,
however, there is little guidance on selection of appropriate values. Applica-
tion of these equations will necessarily involve judgment that can only be
applied on a case-by-case basis.
27
Chapter 2 Contaminant Losses During Dredging
-------
3 Contaminant Losses During
Dredged Material Transport
Background
This section is concerned with contaminant losses during transportation of
dredged material. Transportation methods include pipelines, scows, barges,
and hoppers. Trucks and railroad cars are rarely used. Hopper dredge trans-
port with direct pumpout is often used in the Great Lakes, but is not the most
common form of dredged material transport.
Losses during transport are easier to control than to predict. Transporta-
tion losses are largely due to accidental spills and leaks, events which are very
difficult, if not impossible, to predict. Controls as discussed by Cullinane
et al. (1986) can significantly reduce these losses. Controls are briefly men-
tioned for each form of dredged material transport discussed below.
Spills and leaks account for all the paniculate and dissolved contaminant
losses and a portion of the volatile losses during dredged material transport.
Volatile losses can be predicted and to some extent controlled. The predictive
techniques discussed in this chapter are, therefore, limited to volatile losses.
Prediction of paniculate and dissolved contaminant losses through spills and
leaks is discussed, but no predictive techniques are available.
Losses During Pipeline Transport
Pipeline operations keep dredged material in a closed system until deliv-
ered to a destination. Pipeline operations, therefore, offer the potential for
zero losses during transport of dredged material. However, accidental
releases through pipeline failures and leaks can occur. In addition, dredge
pump outages due to damage by objects entrained in the suction (nuts, bolts,
chain, cables, rocks, etc.) can result in clogged pipelines that have to be disas-
sembled and cleaned. During disassembly and cleaning, losses can occur.
Since pump outages and pipeline failures and leaks are unpredictable, there
are no a priori techniques for predicting contaminant losses during pipeline
28
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
transport of dredged material. Ideally, the losses during pipeline transport
should be zero, but in reality losses are never zero.
During the design stage, planners should carefully consider pipeline routes,
climatic conditions expected, material's corrosion resistance, redundancy of
safety devices (i.e., additional shutoff valves, loop/by-passes, pressure relief
valves), coupling methods, and systems to detect leaks.
Losses During Scow, Barge, and Hopper Transport
Contaminant losses from scows, hoppers, and barges can occur via resus-
pension of sediment as a result of spillage, overflow, and volatilization. The
manner in which dissolved and paniculate contaminants are lost depends on
the type of dredging operation (mechanical or hydraulic) used to fill the trans-
port vessel. Volatile contaminant releases, which also depend on the type of
dredging, are discussed in a later section.
Material condition prior to placement into a scow or barge has a great
impact on what controls planners must consider. Dredged material that has a
high moisture content will require less concern about possible windblown
dust, but will create much more difficult loading and unloading conditions and
will require a greater number of barges. In general, lower material moisture
content is better for handling and control. For purposes of discussing control
mechanisms in barge transport, the dredged material will be assumed to be in
one of two states: freshly dredged material, having a very high water content
and being transported a short distance to an unloading site, or consolidated
(dewatered) dredged material to be barge transported over long distances.
Bucket operations
Since bucket dredging produces dredged material at close to in situ densi-
ties, overflow from a scow or barge can be controlled such that overflow
losses are negligible during transportation. Loading and unloading probably
presents the greatest potential for uncontrolled contaminant releases during
bucket operations. At loading and unloading points, spillage directly in the
water can occur during boom swing between the transport vessel and the
delivery point. Controls can be implemented to significantly reduce or elimi-
nate this type of loss. Techniques for predicting contaminant losses during
unloading operations using buckets are not available.
When volatile or semivolatile contaminants are present in the dredge mate-
rial, open-top transport vessels are a continuous source of volatile emissions
until emptied. Volatilization rates from open-top vessels depend on sediment
volatile chemical concentrations, wind speed, area of exposed dredged mate-
rial, and physical/chemical properties of the contaminants. Predictive
techniques for volatile losses from mechanically dredged sediment during
transport in open-top vessels are described in a later section.
29
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
Hydraulic operations
Scows, barges, and hoppers can be loaded hydraulically as well as
mechanically. In addition, mechanically loaded vessels can be hydraulically
unloaded. Since hydraulic unloading involves pipelines, losses during hydrau-
lic unloading are similar to pipeline losses, that is, uncontrolled spills and
leaks are the major contaminant loss mechanisms. As previously mentioned,
open-top transport vessels may be a continuous source of volatile emissions
until emptied. Predictive techniques for volatile losses from hydraulically
dredged sediment during transport in open-top vessels are described in a later
section.
Hopper dredges are sometimes allowed to pump past overflow in order to
achieve better dredging economics by trapping heavy, coarse-grained materials
and releasing light, fine-grained materials. Hopper overflow is a major source
of contaminant reentry into the environment because contaminants preferen-
tially bind to fine-grain materials. Losses during hopper overflow were not
considered in Chapter 2 and are not considered in this chapter because it
makes little sense to remove contaminated sediment for purposes of remedi-
ation and then put the fine-grained fraction (the fraction containing most of the
contaminants) right back in the water. Hopper dredging is a dredging option
that should be considered for remediation, but overflow is not recommended.
Contaminant losses during direct hopper pumpout to treatment or disposal
facilities are essentially the same as pipeline losses previously discussed. The
major difference is that the pipeline distance for direct hopper pumpout is
significantly less than the distances normally used in hydraulic dredge pipeline
operations. The potential for spills and leaks during transportation is, there-
fore, less for hopper dredging than for pipeline dredging.
Losses During Truck and Rail Transport
Truck and rail transport, not often used to transport dredged material
during navigation dredging, has a higher probability of being used in the
transport of contaminated dredged material during remedial operations. Truck
and/or rail transport may be needed when the destination is not accessible by
water or the transportation distance is longer than the range normally used for
overland pipelines. The types of losses for truck and rail transport include
spillage during loading and unloading operations, spills and leaks during
hauling, and volatile emissions throughout the entire cycle of loading, hauling,
and unloading. Accidental spills and leaks are unpredictable losses that can be
controlled by proper planning. Volatile emissions from open-top trucks and
rail cars are predictable and to some extent controllable. Predictive techniques
for volatile losses from open-top vessels are discussed in a later section.
Loading and unloading operations probably present the greatest potential
for contaminant loss when using truck or rail transport. During loading and
30
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
unloading operations involving buckets, conveyer belts, and slides, there will
be some spillage of dredged material. Loading and unloading sites will
become contaminated by spilled materials unless lined. Undercarriage wash-
ing to prevent contaminated sediment from falling on roadways and railways
will generate rinse water that may require treatment. Truck and railcar clean-
ing will also result in wastewater that may require treatment. Treatment
process trains likely to be considered for these wastewaters include sedimenta-
tion, clarification, carbon adsorption, and biological treatment. Discussion of
losses from treatment processes are covered later in this report.
Regardless of loading method, there will be some spillage of contaminated
materials. Controls suggested for consideration are as follows (Cullinane
et al. 1986):
a. Drainage of water from loading and unloading area into central sump
for periodic removal.
b. Daily removal of spilled material.
c. Specially designed loading ramps to collect spilled material.
d. Use of watertight clamshells for transferring materials from barges into
truck.
Volatile Losses During Dredged Material Transport
Volatilization processes differ for exposed sediment solids and sediment
solids covered by water (Thibodeaux 1989). In vessels filled hydraulically
(scows, barges, hoppers), dredged material solids will be covered by water.
In open-top vessels filled mechanically (scows, barges, trucks, and railroad
cars), the dredged material solids may be exposed. Two predictive tech-
niques, one for exposed dredged material solids and one for dredged material
solids covered by water, are discussed in this section.
Volatilization of chemicals from open-top vessels during dredged material
transportation is essentially independent of vessel type. The surface area of
the vessel, however, is important because volatile emission rates depend on
surface area.
Mechanically dredged sediment
Contaminated sediment that is wet and exposed directly to the atmosphere
is the case that results in the highest instantaneous volatile fluxes because the
pathway for loss is very short (Thibodeaux 1989). The water film covering
exposed solids is very thin and provides little resistance to mass transfer
across the solids-air interface. Thus, most of the resistance to mass transfer
resides in the air side of the solids-air interface. Assuming negligible
31
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
resistance to mass transfer on the sediment side (including water films on
sediment solids), the volatilization rate from an open-top vessel containing
mechanically dredged sediment is given by (Thibodeaux 1989)
R - K
KV,es ~ KOG
PA
(16)
where
RVes = volatile emission rate for chemical A from exposed sediment,
g/cm2 sec
KOG = overall gas-side mass transfer coefficient, cm/sec
Av = surface area of vessel, cm2
Pj = density of air, g/cm3
pA* = partial pressure of chemical A in air that would be in equilibrium
with dredged material, mm Hg
pA = background partial pressure of chemical A in air, mm Hg
P = total atmospheric pressure, mm Hg
Equation 16 is Equation 15 from Thibodeaux (1989) written in terms of partial
pressures. The driving force modeled by Equation 16 is the difference
between the partial pressure of a chemical in the air immediately adjacent to
contaminated dredged material and the partial pressure of the chemical in the
background air. The driving force is maximized when the partial pressure in
the air at the contaminated solids-air interface is maximized. The maximum
partial pressure in the air at the contaminated solids-air interface is the equilib-
rium partial pressure, p*A. Generally, p*A can be determined by Henry's Law
partitioning between dissolved concentrations in the dredged material pore
water and air as follows (Thibodeaux 1979):
= H Cw . H (17)
where H equals Henry's constant, mm Hg t /mg.
If the sediment surface is approximately flat, turbulent boundary layer theory
suggests that overall gas-side mass transfer coefficient is given by
32
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
KOG = 0.036
D
Al
0.8
0.33
(18)
D
Al
where
DA1 = molecular diffusivity of chemical A in air, cm2/sec
Lv = vessel length, cm
Vx = background wind speed, cm/sec
Vj = kinematic viscosity of air, cm2/sec
Uneven surfaces will tend to increase KOG if the surface roughness occurs
within the boundary layer. Large mounds increase surface area and shade
downwind areas (decrease effective surface area), neither of which is a term in
Equation 18.
For long transportation times or for long-term storage before disposal, the
surface of the dredged material will lose both water and contaminant, and
volatilization will slow due to the development of internal mass transfer resis-
tances. Procedures for estimating contaminant volatilization from exposed
dredged material with internal resistances is discussed in Chapter 4.
The equation for estimating volatile losses when open-top vessels are par-
tially filled (loading and unloading) is given below (Thibodeaux 1989)
R,.
2DV - Z
2Z>,,
R
V,es
(19)
where
Dv = effective diameter of vessel, cm
Z = distance from top of vessel to exposed dredged material surface, cm
The term in parenthesis in Equation 19 accounts for the exposed surface being
a distance Z below the top of a vessel with an effective diameter Dv.
Hydraulically dredged sediment
When hydraulically dredged sediment is placed in open-top vessels for
transportation to a destination, the dredged material solids will tend to settle.
Volatilization during transportation of hydraulically dredged sediment in open-
top vessels, therefore, takes place at an air-water interface. Volatilization
Chapter 3 Contaminant Losses During Dredged Material Transport
33
-------
from water surfaces is discussed in Contaminant Losses During Pretreatment
in the section on volatile releases from ponded water.
34
Chapter 3 Contaminant Losses During Dredged Material Transport
-------
4 Contaminant Losses During
Pretreatment
Background
Pretreatment as used in this report is the processing of dredged material for
additional treatment or disposal. Dredged material slurries produced by
hydraulic dredging may require pretreatment to increase the solids content
when treatment technologies designed for low moisture soil, such as thermal
technologies, are to be used (Averett et al. 1990). Because the rate of
dredged material removal and transportation is usually irregular, flow equal-
ization may also be necessary before initiating treatment. Flow equalization
facilities can also serve as a convenient point for dewatering by primary
settling.
Averett et al. (1990) surveyed the applicability of the pretreatment pro-
cesses shown in Table 3 to dredged material. This chapter discusses contami-
nant losses from primary settling and flow equalization facilities.
Table 3
Process Options for the Pretreatment Component1
Dewatering
Belt filter press
Carver-Greenfield evaporation
Centrifugation
Chamber filtration
Evaporation
Gravity thickening
Primary settling (CDF)
Solar evaporation
Subsurface drainage (CDF)
Surface drainage (CDF)
Vacuum filtration
Wick drains (CDF)
Particle Classification
Flotation
Grizzlies
Heavy media separation
Hydraulic classifiers
Hydrocyclones
Impoundment basins (CDF)
Magnetic and electrostatic
separation
Moving screens
Shaking tables
Spiral classifier
Stationary screens
Slurry Injection
Chemical Clarification
Microbe addition
Nutrient addition
1 From Averett et al. (1990).
Chapter 4 Contaminant Losses During Pretreatment
35
-------
Losses During Primary Settling and Flow
Equalization
Primary settling and flow equalization facilities similar in design and oper-
ation to confined disposal facilities (CDFs) (Figure 4) will probably be needed
for hydraulically dredged material. Storage facilities similar to CDFs may
also be needed to stockpile mechanically dredged material for subsequent
treatment. Since primary settling and flow equalization at the beginning of a
treatment process train for dredged material will likely be extensions of exist-
ing CDF technology, techniques that have been developed for estimating
losses from CDFs should be applicable to primary settling and flow equaliza-
tion facilities.
INFLUENT
VOLATILE
EMMISSION
DREDGED \.
MATCBI Ai
AREA FOR SEDIMENTATION
V. * .* *
s^^ ** ^ * J . j * * .. . • • •
WEIR
EFFLUENT
SURFACE
RUNOFF
VOLATILIZATION
UNSATURATED ZONE
PRECIPITATION J
INFILTRATION \
WEIR
SATURATED ZONE :
DREOOEO MATERIAL
EFFLUENT
FOUNDATION
SOILS
LEACHATE
Figure 4. Pretreatment facility schematic with major contaminant migration pathways
(a) during filling and (b) filled
36
Chapter 4 Contaminant Losses During Pretreatment
-------
As shown in Figure 4, the major contaminant loss pathways for pretreat-
ment facilities are effluent, leachate, runoff, and volatilization. Predictive
techniques for estimating contaminant losses along each of these migration
pathways are presented in this section. Discussion of laboratory and field data
for these migration pathways is presented in Losses From Confined Disposal
Facilities on CDF disposal.
Effluent-hydraulic filling
In this section, procedures for estimating effluent contaminant losses during
hydraulic filling of primary settling/flow equalization facilities are discussed.
Treatment technologies that could be applied to the effluent, such as chemical
clarification and carbon adsorption, are discussed in Contaminant Losses
During Effluent and Leachate Treatment.
Data requirements for estimating effluent losses during hydraulic filling are
listed in Table 4. As indicated in Table 4, information on facility design and
influent flow and quality are needed in order to estimate effluent flow and
quality.
Table 4
Data Requirements for Predicting Contaminant Losses During
Hydraulic Filling1
Data Required
Dredge inflow
Influent solids concentration
Influent total contaminant concentrations
Average ponding depth
Hydraulic efficiency factor
Effluent suspended solids concentration
Contaminant dissolved concentrations in
effluent
Fraction of contaminant in effluent sus-
pended solids
Source of Data
Project information, site design
Project information
Bulk chemical analysis of in situ sediment
Project information, site design
Dye tracer study or theoretical retention time
Column settling tests
Modified elutriate test
Modified elutriate test
1 From USAGE (1987).
Influent characteristics. The initial step in any dredging activity is to
estimate the in situ volume of sediment to be dredged. Sediment quantities
are usually determined from channel surveys. Field sampling is required to
characterize the sediment and provide material for laboratory testing. Impor-
tant sediment characteristics that should be determined include water content,
grain-size distribution, Atterberg limits, organic content, specific gravity,
Unified Soil Classification System (USCS) classification, and bulk chemical
concentrations. Although some of this information is not explicitly used to
estimate contaminant losses, prediction of effluent quality is based on facility
design; most of this information is needed to design a primary settling facility.
Palermo, Montgomery, and Poindexter (1978) and USAGE (1987) provide
Chapter 4 Contaminant Losses During Pretreatment
37
-------
guidance on designing primary settling facilities and the data required for
design. Guidance on the collection of sediment samples is provided in the
"ARCS Assessment Guidance Document" (USEPA 1994b).
Influent flow is based on dredge production rates. This type of informa-
tion is usually available in Corps of Engineers District records of dredging
activities. If no data are available, hydraulic pipeline dredge production rates
can be estimated from relationships among solids output, dredge size, pipeline
length, and dredging depth (Palermo, Montgomery, and Poindexter 1978;
USAGE 1987). Figure 5 shows solids production rates for selected pipeline
dredge sizes, pipeline lengths, and dredging depths. For hopper dredges,
disposal rate must be estimated from hopper or barge pump-out rate and travel
time involved. Site-specific records of previous dredging activities are the
best sources for this type of information.
Influent solids concentration will vary with type and size of dredge(s) and
in situ sediment concentration. If data from Corps of Engineers dredging
records are not available, an influent solids concentration of 145 g/f (13 per-
cent by weight) for hydraulic pipeline dredging can be used (Palermo, Mont-
gomery, and Poindexter 1978). This number is based on a number of field
investigations conducted under the DMRP.
Chemical concentrations in the influent can be estimated from bulk chemi-
cal analysis of the in situ sediment and solids concentration of the influent.
Because site water quality has little effect on influent quality, influent contami-
nant concentrations usually reflect dilution of in situ sediment bulk chemical
concentrations. It is sometimes informative to compare site water quality and
effluent quality, but site water quality data are not required for prediction of
effluent quality.
Effluent flow. Effluent flow is approximately equal to influent flow dur-
ing hydraulic filling of sedimentation basins. Initially, there may be some
storage of water in facilities with overflow weirs; however, after the head on
the weir stabilizes, effluent flow is approximately equal to influent flow.
Effluent quality. Effluent quality during hydraulic filling is predicted on
the basis of data from column settling and modified elutriate tests and sedi-
mentation basin design. The modified elutriate test was developed as part of
the LEDO research program to simulate the physicochemical conditions in
CDFs during hydraulic disposal and involves measurement of both dissolved
and total concentrations of contaminants in the elutriate (Palermo 1986). A
separate column settling test is used to predict suspended solids concentration
in effluent for a specific facility design and set of operational conditions.
Results from the modified elutriate and settling column tests are then com-
bined to predict total and dissolved contaminant concentrations in effluent
during hydraulic disposal.
The modified elutriate and companion settling tests when used as described
by Palermo (1986) account for both dissolved and paniculate bound
38
Chapter 4 Contaminant Losses During Pretreatment
-------
O
V)
a
0 4 8 12 16 20
PIPELINE LENGTH, FEET x 1,000
a. DREDGING DEPTH OF 20 FT
0 4 8 12 16 20
PIPELINE LENGTH, FEETx 1,000
c. DREDGING DEPTH OF 40 FT
8 12 16 20
PIPELINE LENGTH, FEETx 1,000
b. DREDGING DEPTH OF 30 FT
24
20
16
12
DREDGE SIZE -
30"
24"
18"
I
0 4 8 12 16 20
PIPELINE LENGTH, FEETx 1,000
d. DREDING DEPTH OF 50 FT
Figure 5. Solids output for selected pipeline dredge sizes, pipeline lengths, and dredging
depths (from U.S. Army Corps of Engineers 1987)
contaminants and the geochemical changes affecting contaminant distribution
between aqueous and dissolved phases during active disposal operations. The
column settling test and facility-specific conditions (surface area, ponding
depth, influent flow, and hydraulic efficiency) are essential parameters for
using the modified elutriate test to predict effluent quality. A flowchart illus-
trating how modified elutriate and column settling tests are used to predict
dissolved and paniculate bound contaminant concentrations in CDF effluent is
shown in Figure 6.
Chapter 4 Contaminant Losses During Pretreatment
39
-------
EVALUATE PERTINENT PROJECT DATA
ON DREDGE AND DISPOSAL AREA
SAMPLE DREDGING SITE
SEDIMENT AND WATER
PERFORM MODIFIED
ELUTRIATE TESTS
PERFORM COLUMN
SETTLING TESTS
ESTIMATE DISSOLVED CONCENTRATION
OF CONTAMINANTS AND FRACTION
IN SUSPENDED SOLIDS
ESTIMATE SUSPENDED SOLIDS
IN DISPOSAL AREA EFFLUENT
ESTIMATE TOTAL CONCENTRATION OF CONTAMINANTS
IN DISPOSAL AREA EFFLUENT
EVALUATE MIXING ZONE AND COMPARE
WITH STANDARDS OR CRITERIA
Figure 6. Steps for predicting effluent quality during hydraulic filling
The procedures shown in Figure 6 have undergone extensive research and
development including field trials. Field studies on maintenance dredging
projects confirmed that the procedures are reliable and usually provide conser-
vative estimates of heavy metal concentrations in effluent (Palermo 1988;
Palermo and Thackston 1988a). Field data for organic contaminants are not
40
Chapter 4 Contaminant Losses During Pretreatment
-------
as extensive as that for metals, but the available field data indicate that the
procedures are also good predictors of organic contaminant concentrations in
CDF effluent during hydraulic filling (Palermo 1986; Palermo 1988; Myers
1991). The modified elutriate and column settling tests are briefly described
below.
The modified elutriate test consists of the following steps (Figure 7):
a. Mixing dredging site sediment and water to the solids concentration
expected in the influent to the facility (discharge from the dredge).
b. Aerating the mixture for 1 hr to simulate the oxidizing conditions
present in primary settling facilities.
c. Settling the mixture for a time equal to the expected or measured mean
retention time of the facility, up to a maximum of 24 hr.
d. Collecting a sample of supernatant for chemical analysis of dissolved
and total contaminant concentrations.
The dissolved concentrations from the test are the predicted dissolved concen-
trations in the effluent. The contaminant concentrations associated with sus-
pended solids are the differences between total contaminant concentrations in
whole water samples and dissolved contaminant concentrations in the filtered
water samples (Equation 20).
r - c
c = mal w (20)
where
Q = solid phase contaminant concentration, mg/kg
Ctotal = whole water contaminant concentration, mg/f
Cw = dissolved contaminant concentration, mg/f
C = suspended solids concentration, kg/£
It should be noted that Cw and Cs in Equation 20 are not necessarily equilib-
rium concentrations. They could be equilibrium concentrations, but equilib-
rium is not a necessary condition in the modified elutriate test.
The column settling test consists of the following steps:
a. Mixing the dredging site sediment and water to the slurry concentration
expected in the influent to the pretreatment or confined disposal
facility.
41
Chapter 4 Contaminant Losses During Pretreatment
-------
WATER FROM
DREDGING SITE
SEDIMENT FROM
DREDGING SITE
MIX SEDIMENT AND WATER TO
EXPECTED INFLUENT CONCENTRATION
( AERATE IN 4 - L CYLINDER
\ FOR 1 HR
CHEMICAL ANALYSIS
TOTAL CONCENTRATION
/ SETTLE FOR EXPECTED MEAN FIELD
\ RETENTION TIME UP TO 24 HR MAXIMUM
EXTRACT SUPERNATANT
SAMPLE AND SPLIT
f CENTRIFUGATION OR
\ 0.45-pm FILTRATION
SUSPENDED SOLIDS
DETERMINATION
CHEMICAL ANALYSIS
DISSOLVED CONCENTRATION
Figure 7. Modified elutriate test procedure
b. Placing the slurry into an 8-in. (20.3-cm) diameter settling column and
allowing it to settle.
c. Taking samples of supernatant water above the sediment-water inter-
face at various time intervals.
d. Analyzing the samples for suspended solids concentrations.
Effluent suspended solids concentration is predicted using the following steps:
a. Developing a relationship of column supernatant suspended solids
concentration versus settling time (Figure 8).
b. Selecting a column supernatant suspended solids concentration corre-
sponding to the expected mean field retention time.
42
Chapter 4 Contaminant Losses During Pretreatment
-------
150
120
90
o
W
60
30
FIELD MEAN RETENTION = 53 HR:
SS
I
60
120
180
TIME, HOURS
240
300
360
Figure 8. Typical plot of supernatant suspended solids concentration versus time for col-
umn settling test (from Palermo and Thackston 1989)
c. Estimating a predicted effluent suspended solids value by adjusting the
value selected in the above step for wind and turbulence under field
conditions using a settling efficiency adjustment factor.
Predicted total contaminant concentrations in effluent during hydraulic filling
are estimated using the following equation:
= C... + C. C. (21)
where CEFFTOT is the total concentration of contaminant in effluent, mg/l.
Detailed information on the development of modified elutriate and column
settling tests including example calculations are provided by Montgomery
(1978); Montgomery, Thackston, and Parker (1983); Palermo (1986); USAGE
(1987); Palermo (1988); Palermo and Thackston (1988a); Palermo and Thack-
ston (1988b); Palermo and Thackston (1988c); Palermo and Thackston (1989);
and Averett, Palermo, and Wade (1988). For a specific dredging project,
hydraulic dredge, and facility design, these procedures have been shown to
reliably predict effluent suspended solids, total contaminant, and dissolved
contaminant concentrations.
When column settling and modified elutriate data are not available, a priori
techniques for estimating effluent quality and mass releases during hydraulic
placement in pretreatment facilities can be used. A priori estimation tech-
niques, by definition, do not require site-specific data. As a result, a priori
estimates are not as reliable as estimates based on site-specific test data.
Chapter 4 Contaminant Losses During Pretreatment
43
-------
Total mass concentration (particulate + dissolved) in effluent during
hydraulic filling can be estimated using the following equation (Myers et al.
1993):
CEFF,TOT = CINF,TOT C1 ~ CEF)
where
CEFFJOT = tota^ concentration of contaminant in effluent, mg/f
CINF.TOT = total concentration of contaminant in influent, mg/f
CEF = contaminant containment efficiency factor for effluent pathway,
dimensionless
The containment efficiency factor (Myers 1991; Myers et al. 1993) is a simple
measure of contaminant mass retention. Palermo (1988) measured CEFs at
five CDFs. The five-site average CEF for metals was 0.986 (98.6-percent
retention). The one site at which PCBs were monitored showed a CEF of
0.99 (99-percent retention) for PCBs. Operation and management of pretreat-
ment facilities for remediation of sediments in the Great Lakes will likely
result in contaminant retention that is at least as good as and probably better
than that measured by Palermo (1988) at CDFs designed and operated for
disposal of dredged materials from maintenance dredging projects.
Dissolved organic contaminant concentrations in effluent can be estimated
using equilibrium partitioning equations described previously in the section on
losses during dredging (Equation 13). Equilibrium partitioning equations
should not be used to estimate dissolved metal concentrations in effluent.
Applications and limitations of equilibrium partitioning equations are discussed
in the next section on leachate losses and in Appendix B.
Effluent-mechanical placement
Influent characteristics. For mechanical dredging and placement, the in
situ water content is a good estimator of the solids content of the dredged
material influent, and bulk chemical analysis of the sediment is a good estima-
tor of influent contaminant concentrations. Influent flow can only be judged
from previous operating records since many site-specific conditions affect the
disposal rate when mechanical dredging and disposal methods are used. For
example, during a 1986 maintenance dredging project in the Chicago River,
dredging was conducted at night. Night work was necessary to minimize
interference with bridge traffic on the many drawbridges that cross the
Chicago River in downtown Chicago. Two barges each containing approxi-
mately 760 m3 (1,000 cu yds) were loaded by a clamshell dredge during the
night and unloaded the following day by clamshell dredge. It took approxi-
mately 3 to 4 hr to unload a barge.
44
Chapter 4 Contaminant Losses During Pretreatment
-------
Effluent quality. In an upland facility, there should be little or no effluent
during mechanical dredging and placement. The small amount of water that
may seep to the surface will have pore water quality. During mechanical
placement of dredged material in nearshore and in-water facilities, water in
the facility before disposal operations begin will be displaced by the dredged
material resulting in a discharge of effluent. Predictive techniques for effluent
quality during mechanical placement of dredged material in nearshore and
in-water facilities that contain water prior to placement of dredged material
are currently unavailable.
Leachate
When contaminated dredged material is placed in a pretreatment facility,
contaminants may be mobilized and transported beyond the facility boundary
by leaching. Leachate is contaminated pore water, and leaching is the combi-
nation of interphase transfer of contaminants from dredged material solids to
pore water and movement of contaminated pore water. Thus, leaching is a
coupling of chemistry and fluid mechanics. Techniques for estimating leach-
ate flow and quality are discussed in this section.
Leachate flow. Leachate flow from dredged material placed in primary
settling facilities and CDFs is produced by four potential water sources:
a. Interstitial water left after primary settling.
b. Rainwater and snowmelt.
c. Offsite groundwater.
d. For in-water facilities, surface water outside the facility.
The predictive technique for estimating leachate flow discussed in this section
accounts for leachate generation associated with the first two water sources.
Application of groundwater models to facilities with leachate generated by
offsite groundwater inflow and techniques for estimating leachate generation
by fluctuating water levels outside a nearshore or in-water facility are dis-
cussed in Losses From Confined Disposal Facilities on CDFs.
After filling is completed, dredged material in a primary settling facility is
initially in a saturated condition (all voids are filled with water). As evapora-
tion and seepage remove water from the voids in the dredged material, the
amount of water stored in the voids and available for gravity drainage
decreases. After some time, usually several years, a quasi-equilibrium is
reached in which water that seeps or evaporates is replenished by infiltration
through the surface. It is not likely that dredged material will be held in
pretreatment facilities long enough for establishment of a quasi-equilibrium.
Leachate flow from primary settling facilities will be time varying and highly
dependent on local climatology, dredged material properties, and facility
45
Chapter 4 Contaminant Losses During Pretreatment
-------
design factors. To predict time-varying leachate flow, all these factors must
be considered.
Preproject estimation of leachate flow, therefore, requires coupled simula-
tion of local weather patterns and surface and subsurface processes governing
leachate generation. The local groundwater regime can be important in evalu-
ating long-term leaching trends at pretreatment facilities. Depending on local
geohydrology, hydraulic conductivity of the dredged material, size of the
facility, and other site-specific factors, such as liners, groundwater flow may
tend to go beneath a pretreatment facility, diverge and spread around it, or
even discharge into it. In most cases, however, leachate flow from a pretreat-
ment facility is governed by the initially saturated condition of the dredged
material, the amount of pore water initially available for gravity drainage, and
the replenishment of water that seeps from the site by rain and snow. In
short, leachate generation at pretreatment facilities is governed by the initial
condition of the dredged material and surface hydrology.
Important climatic parameters include precipitation (rain and snow), tem-
perature, and humidity. Important surface processes include snowmelt, infil-
tration, surface runoff, and evaporation. Important subsurface processes
include evaporation from dredged material voids and flow in vadose and
saturated zones in the dredged material. Important facility design factors
include hydraulic properties of the foundation soils, type of liner (if any), and
type of leachate collection system (if any). Due to the complexity of the
interactions among climatic events, surface hydrologic processes, and subsur-
face hydraulics, there is no one laboratory test capable of predicting leachate
flow.
There is, however, a simulation model available that couples climatic
events, surface hydrologic processes, and subsurface hydraulics that is applica-
ble to dredged material in an upland containment facility. This model is the
Hydrologic Evaluation of Landfill Performance (HELP) computer program
(Schroeder et al. 1988). HELP is a hydrologic water budget model that
accounts for the effects of surface storage, runoff, infiltration, percolation,
evapotranspiration, soil moisture storage, lateral drainage to leachate collec-
tion systems, and percolation through synthetic liners, soil liners, and
composite liners. Local climatology is one of the important components of
hydrologic modeling that the HELP model simulates on a daily basis.
The HELP model was developed by the USEPA to predict the amounts of
seepage, drainage to leachate collection systems, at sanitary landfills. The
model is used in a preproject mode by designers and permits writers to evalu-
ate landfill designs. HELP model features that are particularly useful for
estimating leachate flow are summarized in Table 5. Limitations that apply
when using the HELP model to estimate leachate flow are also summarized in
Table 5,
The HELP model simulates flow through as many as 12 layers with vary-
ing hydraulic properties. The first layer is usually a cap, and the bottom layer
46
Chapter 4 Contaminant Losses During Pretreatment
-------
Table 5
HELP Model Major Features
Advantages
Time varying.
Simulates site-specific climatology on a daily basis using user-supplied data, default data
for 102 cities in the U.S., or a synthetic climatology generator.
Couples vegetative growth, evapotranspiration, surface runoff, unsaturated flow, saturated
flow, and soil moisture storage.
Layers of differing hydraulic properties simulated, including caps, liners, and leachate col-
lection systems.
Includes default climatological and soil property data.
Interactive and runs on desktop computers.
Documented.
Coupling of surface, vadose, and saturated flows experimentally verified (Schroeder and
Peyton 1987).
Limitations
One-dimensional (vertical percolation) except for leachate collection systems.
Assumes bottom is free draining.
is usually a low-permeability barrier soil, although these are not model
requirements. The model is quasi-two-dimensional in that layers can be
defined as lateral drainage or vertical percolation layers. Lateral drainage
layers are appropriate for designs that include a leachate collection system.
Without lateral drainage layers, subsurface flow calculations in the HELP
model are one-dimensional simulations of vertical percolation.
A definition sketch for application of the HELP model to recently filled
primary settling facilities is shown in Figure 9. As shown in Figure 9, the
dikes should be impermeable relative to the hydraulic conductivity of the
dredged material. These conditions are not always met, but when they are,
flow into the foundation soils is primarily vertical. In this case, the physical
system closely matches the HELP model assumptions so that there are few if
any limitations to application of the HELP model.
The general simulation parameters (user-supplied inputs) are listed in
Table 6. The user must specify the number and thickness of each layer.
There are three types of layers in the HELP model as follows: vertical perco-
lation layers, lateral drainage layers, and barrier soil liners. Vertical
percolation layers are layers without a leachate collection system. The
dredged material in a primary settling facility would be classified as a vertical
percolation layer. If there are dredged material layers with different proper-
ties, such as hydraulic conductivity, dredged material layers as needed could
be specified as vertical percolation layers as long as the total number of layers
Chapter 4 Contaminant Losses During Pretreatment
47
-------
PRECIPITATION
EVAPOTRANSPIRATION
i \
DREDGED MATERIAL
. • . . . LEACHATE COLLECTION . .
• FLEXIBLE MEMBRANE LINER
•'.••'•'.•'.••'•'•''•'•' .LEACHATE COLLECTION- . • .' •
FLEXIBLE MEMBRANE LINER
Figure 9. Definition sketch for application of HELP model to primary settling facilities for
dredged material
does not exceed 12. Lateral drainage layers are layers designed to collect
leachate by lateral drainage to collection pipes. Both vertical and lateral
drainage are simulated by the HELP model in lateral drainage layers. A layer
of material design to inhibit percolation is classified as a barrier soil liner. A
layer covered by a flexible membrane liner (FML) is classified as barrier soil
liner with an FML. In addition, the user can select the "active" or "closed"
options. The "active" option will not allow runoff to occur. Excess
48
Chapter 4 Contaminant Losses During Pretreatment
-------
Table 6
General Simulation Parameters for the HELP Model
Facility Design Parameters
Number of layers (1 to 12)
Layer classification as vertical percolation, lateral drainage, or barrier soil liner
Thickness of each layer
Liner presence (yes/no) for barrier soil liners
Open or closed site
Surface area
Climatological Database Choices
Default database (4-year record) for 102 U.S. cities
User supplied database
Synthetic weather generator for 139 U.S. cities
Soil and Dredged Material Properties
Default soil option (yes/no)
Manual soil option
Porosity
Field capacity
Wilting point
Initial water content
Saturated hydraulic conductivity
Other
Evaporative zone depth
Type of vegetative cover
Simulation period (1 to 20 years, depending on climatological database)
Type of output
Daily
Monthly averages
Annual totals
precipitation will pond on the dredged material surface. The "closed" option
will allow runoff.
The user has the choice of using a default climatological database, user-
supplied database, or a synthetic weather generator. The default climatologi-
cal database is a 5-year record (1974 through 1978) for 104 U.S. cities. The
Chapter 4 Contaminant Losses During Pretreatment
49
-------
user can choose to input a climatological database consisting of daily tempera-
ture, precipitation, solar radiation, and other parameters. The HELP model
synthetic weather generator is applicable to 139 U.S. cities. The default soil
database in the HELP model is based on the USCS. The user specifies the
type of soil according to one of 15 possible USCS classifications. There are
also default soil data for two types of barrier soils that may be further speci-
fied as compacted or uncompacted. The user can also specify soil and
dredged material properties for each layer as follows: wilting point, porosity,
saturated hydraulic conductivity, initial water content, and field capacity.
Other model inputs include evaporative zone depth, leaf area index, simu-
lation period, and type of output. The evaporative zone depth is the depth
beginning at the soil cover (or dredged material) surface affected by evapora-
tive drying. The leaf area index is zero for a primary settling facility since
the dredged material will be removed for treatment before vegetation has a
chance to establish. The maximum simulation period is 20 years and depends
on the length of record in the climatological database. Longer periods can be
simulated by restarting the HELP model using water budget information from
the last output.
The types of output data provided by the HELP model when the user
specifies daily output are listed below:
Julian date
Precipitation, inches
Runoff, inches
Evapotranspiration, inches
Head on barrier soil liners, inches
Percolation through barrier soil liners, inches
Lateral drainage from surface of any barrier soil, inches
Water content in evaporative zone, dimensionless
The following types of output are provided when the user specifies monthly
totals:
Precipitation, inches
Runoff, inches
Evapotranspiration, inches
For each layer:
Percolation, inches
Lateral drainage, inches
Monthly average daily head, inches
Monthly standard deviation of daily heads, inches
Annual totals for the parameters listed below are presented for both daily
and monthly output options in three types of units: inches, cubic feet, and
percentage of annual precipitation:
50
Chapter 4 Contaminant Losses During Pretreatment
-------
Precipitation
Runoff
For each layer
Percolation
Lateral drainage
Soil water in storage at beginning of year
Soil water in storage at end of year
Snow water in storage at beginning of year
Snow water in storage at end of year
Annual change in total water storage
Leachate quality. Techniques for predicting leachate quality in primary
settling facilities and CDFs are discussed in this section. Two types of predic-
tive techniques for leachate quality are discussed. The first technique is an
a priori technique, and the second technique involves laboratory leach tests.
Both techniques are based on equilibrium partitioning theory. Application of
this theory to dredged material leaching is described by Hill, Myers, and
Brannon (1988); Myers, Brannon, and Price (1992), Brannon, Myers, and
Tardy (1994).
Equilibrium partitioning as used in this report is a simple representation of
a variety of contaminant interphase transfer processes. The complexity of the
problem is shown in Figure 10. As shown in Figure 10, interphase contami-
nant transfer is a complicated interaction of many elementary processes and
factors affecting these processes. A complete description of all these pro-
cesses, their interactions, and factors affecting these processes is not presently
possible. Instead, a lumped parameter, the equilibrium distribution coeffi-
cient, is used to describe the distribution of contaminant between aqueous and
solid phases.
At equilibrium, the net transfer of contaminant across the solids-water
interface is zero, and the mass of contaminant in each phase is constant, but
not necessarily equal. Thus, only the relative distribution of contaminant
between solid and aqueous phases is needed to predict leachate quality. This
distribution of contaminant mass between solid and aqueous phases is repre-
sented by the equilibrium distribution coefficient defined as follows:
K - L (23)
51
Chapter 4 Contaminant Losses During Pretreatment
-------
SEDIMENT SOLIDS
GEOCHEMICAL PROCESSES
ELEMENTAL PARTITIONING
INTRAPARTICLE PORE PHENOMENA
FACTORS AFFECTING INTERPHASE TRANSFER
REDOX POTENTIAL
IONIC STRENGTH
HYDROGEN ION CONCENTRATION
SEDIMENT ORGANIC CARBON
PORE WATER VELOCITY
PORE WATER
-*- DISSOLUTION
PRECIPITATION
ADSORPTION
DESORPTION
SURFACE COMPLEXATION
•*- SOLUBLE COMPLEXATION
COLLOID
FLOCCULATION/
DEFLOCCULATION
CONVECTION
DISPERSION
PROCESSES OCCURRING IN BOTH PHASES
BIODEGRADATION
BIOTRANSFORMATION
CHEMODEGRADATION
CHEMOTRANSFORMATION
REDOX REACTIONS
ACID-BASE REACTIONS
Figure 10. Interphase transfer processes and factors affecting interphase transfer processes
(from Myers, Brannon, and Price 1992)
52
Chapter 4 Contaminant Losses During Pretreatment
-------
where
Kd = contaminant-specific equilibrium distribution coefficient,
dimensionless
Mcs = mass of contaminant in solid phase, kg
Ms = mass of solids, kg
M^ = mass of contaminant in aqueous phase, kg
Mw = mass of water, kg
The mass fractions in Equation 23 can be replaced with phase contaminant
concentrations without any loss of generality so that Equation 23 becomes
(24)
where
Kd = contaminant-specific equilibrium distribution coefficient, *7kg
Q = contaminant concentration in sediment at equilibrium, mg/kg
Cw = aqueous phase concentration at equilibrium, mg/£
Equations 23 and 24 describe the equilibrium distribution of a single con-
taminant in dredged material; that is, equilibrium distribution coefficients are
contaminant and dredged material specific. In addition, the distribution of
contaminant mass is affected by various factors, such as pH, ionic strength,
redox potential, and sediment organic carbon. Varying these factors during
leaching can shift the equilibrium position of the system and change Kd.
The local equilibrium concept is illustrated in Figure 1 1 . When the rate at
which water moves is slow relative to the rate at which equilibrium is
approached, a local chemical equilibrium exists between the pore water and
the sediment solids. Thus, the local equilibrium assumption implies that as a
parcel of water passes a parcel of dredged material solids, the water and solids
come to chemical equilibrium before the parcel of water moves to contact the
next parcel of dredged material solids. Thus, leachate quality at the surface
can differ from leachate quality at the bottom of a primary settling facility,
while leachate in both locations will be in equilibrium with the dredged mate-
rial solids.
Application of the equilibrium assumption to prediction of leachate quality
in dredged material is based on two arguments: (a) the argument that the
interphase transfer rates affecting leachate quality are fast relative to the
53
Chapter 4 Contaminant Losses During Pretreatment
-------
INCREMENT
C1=(V,/Kd1
INCREMENT
C2 -
INCREMENT
= csn/Kdn
PORE WATER
THE PORE WATER IN EACH INCREMENT COMES TO EQUILIBRIUM
WITH THE SEDIMENT SOLIDS IN THAT INCREMENT
BEFORE MOVING INTO THE NEXT INCREMENT
Figure 11. Illustration of local equilibrium assumption for leaching in a pretreatment facility
volumetric flux of water and (b) the argument that equilibrium-controlled
desorption provides conservative predictions of leachate quality. These argu-
ments are discussed below.
The equilibrium assumption is valid when the seepage velocity is slow
relative to the rate at which contaminants desorb from dredged material solids.
This is a realistic assumption for fine-grain dredged material because seepage
velocities are usually very low due to the low hydraulic conductivity of fine-
grain dredged material. The hydraulic conductivity of fine-grain dredged
material is usually in the range of 10"8 to 10~5 cm/sec. In primary settling
facilities and CDFs, the hydraulic gradient is usually about one, so that pore
water velocities are usually in the range of 10~8 to 10"5 m/sec. Some soil
column studies have indicated that the local equilibrium assumption is valid
for pore water velocities as high as 10"5 cm/sec (Valocchi 1985). Theoreti-
cally, equilibrium-controlled desorption requires an infinitely fast desorption
rate. However, if the critical interphase transfer rates are sufficiently fast, the
equilibrium assumption can yield results indistinguishable from full kinetic
modeling (Jennings and Kirkner 1984; Valocchi 1985; Bahr and Rubin 1987).
54
Chapter 4 Contaminant Losses During Pretreatment
-------
In addition to being a good approximation, the assumption of equilibrium-
controlled desorption is conservative; that is, predictions based on the equilib-
rium assumption will overestimate leachate contaminant concentrations if pore
water velocities are too high for local equilibria to become established. The
equilibrium assumption is conservative because interphase transfer is from the
dredged material solids to the pore water, and equilibrium means that all of
the desorption that can occur has occurred. Thus, for clean water entering
dredged material, pore water contaminant concentrations cannot be higher than
the equilibrium value.
Rearrangement of Equation 24 yields
Cw = -& (25-a)
Equation 25-a uses the bulk sediment contaminant concentration, Cs, and a
contaminant-specific distribution coefficient, Kd, to predict dissolved leachate
contaminant concentration.
Organic contaminants sorb to the humic and fluvic acids that make up
dissolved organic carbon. Since dissolved organic carbon is mobile, dissolved
organic carbon enhances advective transport of contaminants. Equation 25-b
includes a factor to account for facilitated transport by colloidal-bound
contaminant.
Kc Cc) - C*(l *KcCc) (25-b)
where
Cpw — Pore water contaminant concentration, mg/£
Cc = dissolved organic carbon, kg/I
Kc = equilibrium distribution coefficient for partitioning of contaminant
between dissolved organic carbon and water, £/kg
Empirical equations that relate distribution coefficients to sediment organic
carbon and octanol-water partitioning coefficients are available (Karickhoff,
Brown, and Scott 1979; Means et al. 1980; Karickhoff 1981; Schwarzenbach
and Westall 1981; Chiou, Porter, and Schedding 1983; Lyman, Reehl, and
Rosenblatt 1990). These relationships were developed mainly through batch
adsorption tests using soils, sediments, and aquifer materials. The generality
of these relationships for desorption of contaminants from dredged material is
uncertain, but the basic technique is widely accepted. A priori estimation of
55
Chapter 4 Contaminant Losses During Pretreatment
-------
distribution coefficients is described in Appendix B. Caution should be exer-
cised when choosing and using Kd values. If Kd is estimated from empirical
relationships based on sediment organic carbon content and octanol-water
partitioning coefficients, Equation 25-b is recommended. Equation 25-a is
valid, but facilitated transport will not be included. If Kd is determined using
the sequential batch leach test discussed later, Equation 25-a should be used
because the Kd obtained from this test included facilitated transport. Equa-
tion 25-b should not be used in this case.
An example a priori prediction of organic chemical concentrations in
dredged material leachate is presented in Table 7. The estimates provided in
Table 7 are sediment and contaminant specific. The predictions are sediment
specific because the Cs values used in the predictions are for a sediment from
Norfolk, VA, and the Kd values are based, in part, on the organic carbon
content of that sediment. The predictions are contaminant specific because the
octanol-water partitioning coefficients used to calculate Kd values are contami-
nant specific.
Table 7
A Priori Prediction of Selected Organic Chemical Concentrations in
Dredged Material Leachate From Norfolk, VA
Organic Contaminant
p,p-DDD
p,p-DDE
p.p-DDT
Heptachlor
Dieldrin
Endosulfan sulfate
Endrin
Endrin Aldehyde
Heptachlor Epoxide
Methoxychlor
C,. mg/kg
0.0004
0.0022
0.0012
0.0022
0.0007
0.0014
0.0003
0.001 1
0.0007
0.0017
Cw. A/g/f
7.2 E-07
6.9 E-06
6.8 E-05
0.0025
0.0057
0.0571
0.0004
7.9 E-06
0.0440
0.0003
Kd, tlkg
55,770
3.2 E + 05
17,600
869
123
24.5
807
1.4 E + 05
15.9
5,794
Note: From Palermo et at. (1993).
Equilibrium partitioning theory with some modification can also be used to
develop a priori predictions of metal concentrations in dredged material leach-
ate (Palermo et al. 1993). The theoretical and experimental basis for a priori
estimation of metal pore water concentrations is not as well developed as that
for organic contaminants. The basic approach for metals is the same as the
approach for organic contaminants except that Equation 25-a as stated is not
applicable to metals. Equation 25-a is not applicable because the total metal
concentration in the dredged material solids is not leachable (Environmental
Laboratory 1987).
56
Chapter 4 Contaminant Losses During Pretreatment
-------
A significant fraction of the total metal concentration in sediments is in geo-
chemical phases that are not mobilized by aqueous extraction (Brannon et al.
1976; Steneker, Van Der Sloot, and Das 1988).
Modification of Equation 25-a for the leachable metal concentration pro-
vides a method for estimating pore water metal concentrations. Assuming a
modification of the equilibrium approach discussed previously applies, metal
pore water concentration is given by
r -
w
(26)
where CsL is the leachable metal concentration in the dredged material solids
(mg/kg).
Empirical relationships for estimating the water leachable concentration,
CsL and the distribution coefficient, Kd, for metals are not available. These
parameters are sediment specific, as well as metal specific. They are affected
by a variety of factors including oxidation-reduction potential, pH, and
organic carbon, sulfur, iron, and salt contents of the sediment. For these
reasons, Kd and CsL are difficult to estimate a priori.
Data from Brannon et al. (1976), Environmental Laboratory (1987),
Brannon, Myers, and Price (1992) on leachable metal fractions in three fresh-
water sediments are presented in Table 8. As indicated in Table 8, between
Table 8
Percent Leachable Metal Concentrations in Selected Sediments
Metal
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Sediment
1
0.34
<0.01
*
<0.01
<0.01
*
0.87
2
6.5
5.2
*
0.55
2.4
1.3
3.0
3
1.37
0.40
0.17
*
*
0.33
0.27
1 Ashtabula Harbor, Ohio; sum of interstitial and exchangeable phases, from Brannon et al.
(1976).
2 Hamlet City Lake, North Carolina; total extracted in anaerobic sequential batch leach
test, from Brannon, Myers, and Price (1992).
3 Indiana Harbor, Indiana; total extracted in anaerobic sequential batch leach test, from
Environmental Laboratory (1987).
* No data.
Chapter 4 Contaminant Losses During Pretreatment
57
-------
about 0.3 and 7 percent of the total arsenic, about 0.01 and 6 percent of the
total cadmium, and 0.2 and 3 percent of the total zinc in these freshwater
sediments were leachable. These ranges in leachable metal fractions can be
used to estimate ranges of CsL for metals in freshwater sediments. The leach-
able concentration is given by multiplying the bulk sediment metal concentra-
tion by the percent leachable divided by 100.
Data on other metals is too limited to provide guidance on estimating
leachable fractions in sediments. Mercury was investigated by Environmental
Laboratory (1987) and Palermo et al. (1989), but detectable amounts did not
leach in sequential batch leach tests. Other studies have also shown that very
little of the mercury in sediments is mobile (Brannon, Plumb, and Smith
1980).
Distribution coefficients are also needed to estimate pore water metal con-
centrations. Anaerobic sequential batch leach data from Environmental Labo-
ratory (1987), Palermo et al. (1989), and Myers and Brannon (1988) indicated
Kd values for metals range from 2 to 90 t/kg, depending on the metal and the
sediment. Conservative estimates are obtained when high values of Kd are
avoided, that is, the lower end of the range in expected Kd values is used.
For conservative estimation of metal pore water concentrations, a range of Kd
values between 3 and 10 £/kg is recommended.
Since specific values for the variables Csl and Kd are not known a priori, a
range of metal pore water concentrations should be estimated. Figure 12 is an
example of the type of concentration envelope that can be developed using a
range of values for CsL and Kd. In Figure 12, arsenic concentrations in
leachate for various CsL and Kd values are shown as a concentration envelope
bounded by 3 < Kd < 10 and 0.005 < leachable fraction < 0.1, where the
leachable fraction (CsL/Cs) is the percent leachable divided by 100. This fig-
ure is not a generic figure since CsL is required in order to calculate a leachate
concentration. Figure 12 is specifically for Cs = 4.2 mg/kg.
Predictions with less uncertainty than the a priori predictions discussed
above can be made if process descriptors, such as distribution coefficients, are
determined experimentally. Currently, USAGE has a research activity within
the LEDO program at WES that is developing laboratory procedures for
investigating interphase transfer processes, testing alternative formulations of
interphase process mathematics, and quantification of interphase process
descriptors. The basic approach of the LEDO leachate research is a semi-
empirical approach that uses a theoretical framework based on mass transport
theory (Figure 13) to guide experimental design and data interpretation (Hill,
Myers, and Brannon 1988). The theoretical framework couples mass trans-
port theory with both batch and column testing in an integrated approach
(Figure 14) (Louisiana Water Resources Research Institute 1990). In the
integrated approach, process descriptors from batch tests, such as distribution
coefficients, are used to predict column elution curves. If predicted and
observed elution histories agree, the conclusion may be reached that the pro-
cesses governing transfer of contaminants from dredged material solids to
CO
Chapter 4 Contaminant Losses During Pretreatment
-------
0.20
002
0.04 0.06
LEACHABLE FRACTION
008
0.10
0.12
Figure 12. Predicted arsenic concentration in leachate for Cs = 4.2 mg/kg and 31 /kg <
Kd < 10f/kg and 0.005 < CsL/Cs < 0.1 (DWL = drinking water limit)
water have been adequately described. Once interphase transfer has been
adequately described, contaminant migration by leaching can be evaluated for
the flow conditions that apply in the field.
Laboratory procedures for both batch and column tests are under develop-
ment (Myers, Brannon, and Price 1992). The batch test involves sequential
leaching of sediment solids in a quick and relatively easy procedure that pro-
vides quantitative interphase transfer process descriptors.
The sequential batch leach procedure used to investigate sediments and
dredged material (Myers and Brannon 1988) is presented below:
STEP 1: Load sediment into appropriate centrifuge tubes and add suffi-
cient deoxygenated distilled-deionized water to each tube to bring
final water-to-sediment ratio to 4:1 by weight (dry sediment
solids). All operations should be conducted in a glove box
under a nitrogen atmosphere.
STEP 2: Shake or tumble tubes for 24 hr.
STEP 3: Centrifuge for 30 min at 6500 x g for organics and 9000 x g
for metals.
Chapter 4 Contaminant Losses During Pretreatment
59
-------
UASS FLUX
IH
UASS FLUX
our
a?
5 = - If
e
(27)
(28)
where
Dp = dispersion coefficient, cm2/sec
z = distance along main axis of flow, cm
v = average pore water velocity, cm/sec
Cw = aqueous phase contaminant concentration, mg/f
5 = interphase contaminant transfer, mg/£ sec
t = time, sec
pb = bulk density, kg/f
e = porosity, dimensionless
Cs = solid phase contaminant concentration, mg/kg
Figure 13. Mathematical model of dredged material leaching (from Hill, Myers, and
Brannon 1988)
STEP 4: Filter leachate through 0.45-u.m membrane filters for metals.
Filter leachate through a Whatman GD/F glass-fiber prefilter
followed by a Gelman AE glass-fiber filter of 1.0-^m nominal
pore size for organics.
STEP 5: Preserve and store samples in the dark at 4 °C until analyzed.
STEP 6: Return to Step 2 after replacing leachate removed in Step 4 with
fresh deoxygenated distilled-deionized water. Repeat the entire
procedure desired number of complete cycles.
Research to date has included investigations of factors affecting leachate
quality, such as liquid-solids ratio and the shake time required to reach
60
Chapter 4 Contaminant Losses During Pretreatment
-------
CONDUCT
SEQUENTIAL BATCH
LEACH TESTS
CONDUCT
STANDARD SOILS
TESTS
CONDUCT
CONTINUOUS-FLOW COLUMN
LEACH TESTS
FORMULATE
SOURCE TERM FOR
EACH CONTAMINANT
OF INTEREST
OBTAIN TRACE OF
CONTAMINANT CONCENTRATION
VERSUS VOLUME OF
LEACHATE PRODUCED
PREDICT EFFLUENT CONCENTRATIONS
FROM CONTINUOUS-FLOW COLUMN BY
SUBTITUTING FLOW AND CHEMISTRY
PARAMETERS
COMPARE PREDICTED
CURVE TO OBSERVED
CURVE
Figure 14. Integrated approach for examining interphase mass transfer (from Louisiana
Water Resources Research Institute 1990)
steady-state leachate concentrations. Results indicate that a four-to-one ratio
of water-to-solids by weight (dry sediment solids) is best, and 24 hr of shak-
ing time is sufficient to achieve steady-state conditions during batch leaching
of sediments (Brannon et al. 1989; Brannon, Myers, and Price 1990; Myers,
Brannon, and Price 1992).
Sequential batch leach tests were used in three major dredged material
disposal alternative evaluations (Environmental Laboratory 1987; Myers and
Brannon 1988; Palermo et al. 1989) to determine how the equilibrium solid
phase contaminant concentration (Cs) was related to the equilibrium aqueous
phase contaminant concentration (Cw) during leaching. A relationship between
Cs and Cw is needed in order to evaluate the source term S in Equation 27.
The source term 5 is obtained from the chain rule as follows:
Chapter 4 Contaminant Losses During Pretreatment
61
-------
c _ _ " s - _ o s w ("29)
T ~aT ~ T ac^ ~aT
Sequential batch leach tests provide the information needed to evaluate
8Cs/dCw.
By sequentially leaching an aliquot of sediment, a table of solid phase
contaminant concentration (Cs) versus aqueous phase contaminant concentra-
tion can be developed and plotted (successive batches have differing Cs and Cw
concentrations). A plot of Cs versus Cw yields a desorption isotherm, the
shape of which indicates the type of desorption. Several types of desorption
isotherms have been observed in sequential batch leaching of sediments (Envi-
ronmental Laboratory 1987; Myers and Brannon 1988; Palermo et al. 1989;
Myers, Brannon, and Price 1992).
The desorption isotherms shown in Figure 15 are typical for metals in
freshwater sediments. A key feature of these desorption isotherms is the
constant slope. The slope is the distribution coefficient, Kd, and it can be
shown that dCs/dCw = Kd. As previously discussed, Kd's obtained from
sequential batch leach tests do not need an adjustment to account for facilitated
transport. In this case, the source term formulation developed using Equa-
tion 29 is relatively simple, and when Equation 27 is solved, predicted metal
concentrations in the leachate decrease as the dredged material solids are
leached by percolating rainwater. This monotonic decrease in aqueous phase
contaminant concentration as the solid phase contaminant concentration
decreases is a characteristic of classical desorption processes.
A commonly observed feature of desorption isotherms for metals in fresh-
water sediments is that the isotherm does not go through the origin. The
intercept is the amount of metal in geochemical phases that is resistant to
aqueous leaching. The difference between Cs and the intercept is equivalent to
the CsL discussed previously. Accurate measurement of CsL is important
because the initial metal pore water concentration needed to set the initial
condition for Equation 27 is calculated using Equation 25(a or b) for organics
and Equation 26 for metals.
Progress in developing a column leach test as a laboratory-scale physical
model of contaminant leaching from dredged material has been slower than
the development of sequential batch leach tests (Myers, Gambrell, and Tittle-
baum 1991; Myers, Brannon, and Price, 1992). Problems with the time
required to run column leach tests and the potential for sample deterioration
during extended sample collection periods have been encountered. An
improved column leaching apparatus has been designed (Figure 16) and is
being used in current column leaching studies (Myers, Brannon, and Price
1992). The new column design increases the number of pore volumes that
can be eluted in a given period of time, minimizes wall effects, and provides
improvements in flow delivery and control.
Chapter 4 Contaminant Losses During Pretreatment
-------
4130
4125
g 4120
g
tc
4115
ZINC
" 4110
o 0
o
t-
0.25 0.50 0.75 1.00
1.25 1.50
o
Ul
V)
Ul
5
in
5
LJ
Ul
O
20.1
20.0
19.9
19.8
19.7
I I I
CADMIUM
LEGEND
ANAEROBIC LEACHING
0 0.002 0.004 0.006 0.008 0.010
AVERAGE STEADY STATE LEACHATE CONCENTRATION,
0.012
Figure 15. Desorption isotherms for zinc and cadmium in Indiana Harbor
sediment (Environmental Laboratory 1987)
Elution curves obtained from column leach tests generally follow the trends
indicated in sequential batch leach tests, although the sequential batch leach
test usually overpredicts contaminant concentrations in column leachates
(Environmental Laboratory 1987; Myers and Brannon 1988; Palermo et al.
1989). An example is shown in Figure 17. Several explanations for differ-
ences in predicted and observed contaminant concentrations in column leachate
are possible, but no single explanation satisfactorily explains all the informa-
tion available (Myers and Brannon 1988). Four explanations that have been
considered are listed below:
Chapter 4 Contaminant Losses During Pretreatment
63
-------
WATER
FROM
RESERVOIR
'/<" STAINLESS STEEL TUBING (OUTLET)
i/2"x '/4" COMPRESSION FITTING
3/g"NUT
%"ALL THREAD ROD
TOP PLATE
STAINLESS STEEL
TUBING (INLET)
0-R1NG SEAL
SINTERED STAINLESS
STEEL DISTRIBUTION
DISK (OJ875"x 10" 01 AM)
SEDIMENT CHAMBER
SINTERED STAINLESS
STEEL DISTRIBUTION
DISK (OJ875"X IO"DIAM)
0-RING SEAL
BASE PLATE
Figure 16. Schematic of improved column leaching apparatus for sediments
and dredged material
64
Chapter 4 Contaminant Losses During Pretreatment
-------
0.0006
0.0005
^ 0.0004
E
3 0.0003
0.0002
0.0001
•PREDICTED
W/DESORPT10H
PREDICTED
W/0 DESORPTION
PERUEAUETER N0.4
PERUEAUETERS
AND 6
PERUEAUETER H0.4
PERUEAUETER N0.6
PERUEAUETER N0.5
t I
PORE VOLUME
Figure 17. Total PCB concentrations in anaerobic column leachate for
Indiana Harbor sediment (from Environmental Laboratory 1987)
a. Short-circuiting in the columns dilutes leachate with clean water.
b. Desorption in the columns is not equilibrium controlled.
c. Contaminant losses are not properly accounted for in collection vessels.
d. Particle disaggregation in batch tests leads to underestimation of distri-
bution coefficients.
Research aimed at determining the cause or causes for the tendency of batch
data to overpredict column data is continuing.
Because the equilibrium assumption used in designing the sequential batch
leach test is a conservative assumption for contaminant desorption, and there
are data from three studies indicating the sequential batch leach test to be a
conservative predictor, the sequential batch leach procedure discussed is the
recommended laboratory leach test for predicting dredged material leachate
quality for freshwater sediments. Until the sequential batch leach test is fully
developed and verified, column leaching and application of the integrated
approach is also recommended. Additional discussion of dredged material
leachate quality prediction including review of available field data is presented
in Losses From Confined Disposal Facilities in the section on leachate.
Chapter 4 Contaminant Losses During Pretreatment
65
-------
Runoff
Runoff is not likely to be a significant contaminant loss pathway during
pretreatment in primary settling facilities and flow equalization facilities that
include engineering controls for runoff. During filling operations, water
added by precipitation will become a minor component of the effluent flow.
Contaminant losses associated with effluent have been previously discussed.
After filling and while dredged material is being held for treatment or dis-
posal, runoff will be a stochastic event that is low volume relative to effluent
during hydraulic filling (a steady event). Runoff can be controlled by ponding
water and allowing it to evaporate. It is, therefore, anticipated that engineer-
ing controls for containment of runoff will be implemented. If, however,
pretreatment facilities are designed and operated such that runoff is not con-
trolled, runoff will carry contaminants out of the facility. If necessary, the
techniques discussed in Losses From Confined Disposal Facilities in the sec-
tion on runoff can be applied to estimate contaminant losses in runoff during
pretreatment.
Volatilization
Volatilization is the movement of a chemical into the air from a liquid
surface. Volatilization from dredged material solids involves desorption
through a water film covering the solids and then from the water to the air.
Because chemicals must enter the water phase before they can volatilize from
dredged material, the tendency of a chemical to volatilize from dredged mate-
rial can be generally related to the Henry's constant. Henry's constant is the
equilibrium distribution of a volatile chemical between air and water if true
solutions exists in both phases (Thibodeaux 1979). There are various ways to
express Henry's constant (Thibodeaux 1979). Two commonly used definitions
that yield dimensionless Henry's constants are given below.
H = L (30)
(31)
where
Ca = dissolved concentration of chemical A in air, g/cm3
Cw = dissolved concentration of chemical A in water, g/cm3
H = Henry's constant, dimensionless
66
Chapter 4 Contaminant Losses During Pretreatment
-------
MA = molecular weight of chemical A, g/mole
p*A = pure component vapor pressure of chemical A, atm
R = universal gas constant, 82.1 atm cm3/mol K
T = temperature, K
C = solubility of chemical A in water, g/cm3
Henry's constant and, therefore, volatilization tendency depend on aqueous
solubility, vapor pressure, and molecular weight. Chemicals with high
Henry's constant will tend to volatilize while chemicals with low Henry's
constant will tend to dissolve in water. As indicated by Equation 30, Henry's
constant is directly proportional to vapor pressure and inversely proportional
to aqueous solubility. Chemicals with similar vapor pressures but different
aqueous solubilities will have different volatilization tendencies. For example,
the vapor pressures for lindane and Aroclor 1260 are 1.2 x 10"8 and 5.3 x
10"8 atm, respectively; but the Henry's constant for lindane is only 2.2 x
10'8, while the Henry's constant for Aroclor 1260 is 0.3 (Thomas 1990a).
Although the vapor pressures for both chemicals are very low, the Henry's
constants differ by four orders of magnitude due to differences in aqueous
solubility. The aqueous solubility of lindane and Aroclor 1260 are 7.3 and
2.7 x 10"3 g/cm3, respectively (Thomas 1990a). This example shows that
vapor pressure is not a good indicator of volatilization tendency from water.
The actual direction of chemical movement across the air-water interface
depends on chemical concentrations in aqueous and air phases and Henry's
constant. The transfer rate (absorption for transfer to water and volatilization
for transfer to air) depends on wind-induced turbulence at the air-water
interface.
Theoretical chemodynamic models for volatile emission rates from dredged
material were described by Thibodeaux (1989). Thibodeaux (1989) identified
four emission locals, each with its own sources and external factors affecting
emission rates. These four locales were as follows:
a. Dredged material transportation devices.
b. Ponded dredged material.
c. Exposed sediment.
d. Vegetation covered dredged material.
Locales b through d are shown in Figure 18. The first locale, volatile losses
during transportation, was discussed previously. The last locale is not appli-
cable to pretreatment facilities because it is anticipated that dredged material
will be removed for treatment or disposal before vegetation can be established.
This section, therefore, discusses volatile losses from two pretreatment
67
Chapter 4 Contaminant Losses During Pretreatment
-------
PLANT-COVERED
SEDIMENT
PONDED
WATER
WEIR
EFFLUENT
Figure 18. Volatilization locales for a CDF
volatilization locales, losses from ponded dredged material and exposed
dredged material solids.
Locale b - ponded dredged material. Dredged material slurries pumped
to primary settling facilities or CDFs undergo sedimentation, resulting in a
thickened deposit of settled material overlain by clarified supernatant (Fig-
ure 4a). Thus, the ponded dredged material locale is characterized by water
containing contaminated suspended solids and a thickened bottom deposit of
dredged material. The volatilization pathway in this case involves desorption
from the contaminated suspended solids followed by transport through the air-
water interface.
The bottom deposit is not part of the pathway because suspended solids
control dissolved contaminant concentrations, and it is dissolved chemicals
that volatilize. While bottom deposits can contribute to dissolved contaminant
concentrations, the contribution from bottom deposits is not important until
the suspended solids concentration becomes negligible. In a primary settling
facility, there is a continuous flux of suspended solids through the water col-
umn while dredged material is being pumped in. Diffusion from bottom
deposits is, therefore, unimportant relative to desorption from suspended
solids in controlling dissolved contaminant concentrations in primary settling
facilities.
The model equation for volatilization from the ponded dredged material
locale is given below (Thibodeaux 1989)
AL = K.
OL
- C)
(32)
68
Chapter 4 Contaminant Losses During Pretreatment
-------
where
Nw = flux through air-water interface, g/cm2 sec
KOL = overall liquid phase mass transfer coefficient, cm/sec
Cw = dissolved contaminant concentration, g/cm3
C*w = hypothetical dissolved chemical concentration in equilibrium with
background air, g/cm3
The dissolved contaminant concentration, Cw, can be estimated using Equa-
tion 25-a, or data on dissolved contaminant concentrations from the modified
elutriate test can be used. The facilitated transport factor (Equation 25-b)
should not be included because contaminants sorbed to colloidal organic mat-
ter must desorb before they can volatilize. For primary settling facilities, the
ponded water area is known and the suspended solids can be predicted using
the column settling tests previously discussed on losses for the effluent path-
way. Equation 32 is applicable when the dissolved contaminant concentration
is constant. Since volatilization continuously removes chemical mass from the
dissolved phase, there is an implicit assumption for application of Equation 32
that either volatilization is so small that it does not affect dissolved chemical
concentrations or there is a source(s) of chemical that replenishes the dissolved
chemical mass as fast as it volatilizes. The effect that volatilization has on
dissolved chemical concentrations depends on physical and chemical properties
of the chemical of interest and site conditions. For these reasons, the relative
significance of volatilization as a process affecting dissolved concentrations
cannot be evaluated without applying a fate and transport model that simulates
all the important processes. In primary settling facilities, however, there are
two sources that can replenish chemical mass lost through volatilization.
First, chemical is being continuously added in dissolved form by disposal
operations. Second, there is a continuous solids flux through the water col-
umn that through partitioning processes tends to maintain constant dissolved
chemical concentrations. For these reasons, the assumption of a constant
dissolved chemical concentration is probably a good approximation of the field
condition. It is also a conservative assumption since the gradient driving the
volatilization process is not allowed to decrease.
Equation 32 has not been field verified for dredged material in pretreat-
ment facilities or CDFs. The equation is, however, widely accepted and has
been verified for volatile chemical emissions from various water bodies and
waste impoundments (Liss and Slater 1974; Billing 1977; Thibodeaux 1979;
Thibodeaux, Parker, and Heck 1984). Probably the largest source of error in
Equation 32 is estimation of the overall liquid phase mass transfer coefficient.
The overall liquid phase mass transfer coefficient depends on a variety of
variable environmental and hydrodynamic factors that are difficult to quantify.
Lunney, Springer, and Thibodeaux (1985) correlated overall liquid phase mass
69
Chapter 4 Contaminant Losses During Pretreatment
-------
transfer coefficients to wind speed and molecular diffusivity in water. Their
correlation is presented below.
KOL - 19.6 V?3 DA661 (33)
where
KOL = over-all liquid phase mass transfer coefficient, cm/hr
Vx = wind speed, mph
DA = molecular diffusivity of chemical A in water, cm2/sec
Other empirical equations are available for estimating KOL, but the Lunney,
Springer, and Thibodeaux (1985) equation is one of the most widely used
equations. If the molecular diffusivity in water is not known, it can be esti-
mated using Oldham's law as follows (Thibodeaux 1979):
DA
DB
0.6
(34)
M
A
where
A = chemical of unknown molecular diffusivity
B = model chemical of known molecular diffusivity
DA = molecular diffusivity of chemical A in water, cm2/sec
DB = molecular diffusivity of chemical B in water, cm2/sec
MB = molecular weight of chemical B, g/mole
MA = molecular weight of chemical A, g/mole
Equation 33 is an empirical model that lumps chemical property and environ-
mental variables into just two parameters, wind speed and aqueous diffusivity.
Since there are no field verification data for Equation 33 at dredged material
pretreatment and disposal facilities, the range of error is not known. It is
estimated that Equation 33 provides KOL values within an order of magnitude.
Part of the potential error is associated with selecting an average wind speed
to represent a range of wind speeds over some period of time.
Chapter 4 Contaminant Losses During Pretreatment
-------
Thomas (1990a) describes some alternative techniques for estimating the
overall liquid phase mass transfer coefficient that are based on two-resistance
theory as follows (Liss and Slater 1974; Thibodeaux 1979):
K.
OL
K,
HKr
(35)
where
KL = liquid-side mass transfer coefficient, cm/sec
KG = gas-side mass transfer coefficient, cm/sec
Although Equation 35 is a theoretical equation, estimation of KG and KL is
highly empirical. Thomas (1990a) suggests using Southworth's correlations
for volatilization of polynuclear aromatic hydrocarbons to estimate KG and KL
as follows:
KG - 0.32 (Vx + Vcurr)
(36)
where
KG = cm/sec
Vx = wind speed, m/sec
Vcun = water velocity, m/sec
For wind speeds less than 1.9 m/sec, KL in cm/sec in given by
KL = 0.0065
^0.969
'curr
Z0.673
32
(37)
where Z is water depth in meters, KL in cm/sec. For wind speeds greater than
1.9 m/sec and less than 5 m/sec,
Chapter 4 Contaminant Losses During Pretreatment
71
-------
KL = 0.0065
yO.969
curr
£0.673
32 o.526(v;")
(38)
\Vhen there exists no mean advective current in a CDF, wind-driven cur-
rents are of the order of 3 percent of wind speed, assuming continuity of shear
stresses at the air-water interface. Thus, Vcurr in Equations 36-38 can be
replaced with 3 percent of the wind speed.
There are numerous empirical equations from stream reaeration studies that
could also be adapted for estimating volatile emissions. Since the only con-
sensus about these equations is that no one equation is superior for modeling
reaeration, these equations are not discussed. It is recognized, however, that
there are other estimation techniques available for mass transfer coefficients
and that most of these techniques give approximately equivalent results.
Thomas (1990a) also discusses using rule-of-thumb values for KG and KL
when making the type of a priori estimates discussed in this report. These
rule-of-thumb values are presented in Table 9.
Table 9
Rule-of-Thumb Values for Liquid- and Gas-Side Mass Transfer
Coefficients (cm/hr)
Vx < 3 m/sec
3 m/sec < Vx < 10 m/sec
Vx > 1 0 m/sec
Sea Surface Conditions
V
3
5-30
<70
*• 2
"o
—
-
--
Kc = 3,000 (18/M,,)1'2
Note: Vx = Windspeed; MA = Molecular weight of contaminant.
1 From Cohen, Cocchio, and Mackay (1978) as cited by Thomas (1990a).
2 Thomas (1990a).
The recommended estimation technique for KOL is Equation 33 followed by a
check against Equation 35 using values from Table 9 for KG and KL. If the
value predicted by Equation 33 is substantially lower than the value predicted
by Equation 35 using data from Table 9, an estimate should be made using
Equations 35-38. If the value predicted by Equation 33 is within a factor of 3
of the value predicted by Equations 35-38, either value is appropriate. If the
two predictions differ by more than a factor of 3, judgment has to be used.
The alternatives are as follows: (a) select the one that seems most appropri-
ate, (b) select the highest value (conservative approach), (c) use the value
predicted by Equation 35 using data from Table 9, or (d) take the average of
all the estimates.
72
Chapter 4 Contaminant Losses During Pretreatment
-------
In view of the lack of field data on volatilization from dredged material
pretreatment and disposal facilities, it is not possible to determine which tech-
nique is the most accurate for estimating mass transfer coefficients. The
correlations in Equations 36-38 were developed, however, for very similar
situations of evaporation from surface impondments. For this reason, alterna-
tive predictive techniques including a rule-of-thumb approach were described
above. The information from the literature suggests that the techniques dis-
cussed in this report should be accurate to within an order of magnitude
(Thomas 1990a).
Locale C - exposed sediment. This volatilization locale is characterized
by sediment that is exposed directly to air and void of vegetative or other
cover. Exposed sediment is probably the most significant of the four volatil-
ization locales as a source of volatile emissions (Thibodeaux 1989). Exposed
sediment will be a source of volatile emissions during various stages of pre-
treatment and flow equalization as follows:
a. The delta formed during primary settling of dredged material slurries
(Figure 4a).
b. The dredged material in filled primary settling facilities after ponded
water is drawn off (Figure 4b).
c. The delta formed during mechanical placement of dredged material in
in-water or nearshore flow equalization facilities.
d. The dredged material in upland flow equalization facilities for mechani-
cally dredged material.
The rate at which chemicals volatilize from exposed sediment is affected by
many factors. Geotechnical properties such as porosity and water content,
chemical factors such as water and air diffusivities, and environmental factors
such as wind speed and relative humidity all affect volatilization rates. In
addition, processes such as air-water-solids chemical partitioning, diffusion of
thermal energy, evaporation of water, and desiccation cracking of the sedi-
ment can have pronounced impacts on volatile emission rates for exposed
sediment. Complete mathematical coupling of all these processes and the
factors affecting these processes into a model equation(s) would lead to a very
complex model requiring site-specific data that are usually unavailable. For
this reason, the vignette models proposed by Thibodeaux (1989) are recom-
mended for a priori prediction.
Dredged material begins evaporative drying and volatile chemical emission
as soon as it is exposed to air. Initially, the chemical emission rate is affected
by gas-side resistance. The top microlayer quickly becomes depleted of vola-
tile chemicals (and water), so that, continuing losses of volatile chemicals
come from the pore spaces within the dredged material. At this point, the
emission process is transient and changes from being gas-side resistance
73
Chapter 4 Contaminant Losses During Pretreatment
-------
controlled to dredged material-side vapor diffusion controlled. The overall
process is modeled by Equation 39 below (Thibodeaux 1989).
1000
7T t
D
a3
e +
Pb
(39)
74
where
ne = instantaneous flux of chemical A through the dredged material-air
interface at time t, mg/cm2 sec
H = Henry's constant, dimensionless
Kd = contaminant specific equilibrium distribution coefficient, cm3/g
Cm = background concentration of chemical A in air at dredged material-
air interface, mg/cm3
TT = 3.14159 ....
t = time since initial exposure, sec
DA3 = effective diffusivity of chemical A in the dredged material pores,
cm2/sec
€j = air-filled porosity, dimensionless
pb = bulk density, g/cm3
KG = gas side mass transfer coefficient, cm/sec
Equation 39 is an idealized diffusion model that describes chemical move-
ment in the unsaturated zone near the air-dredged material interface. The
emission pathways modeled include surface depletion, desorption from particle
surfaces into a water film surrounding the particle surfaces (hence, the appear-
ance of Kd), desorption from the water film into the pore gas (hence, the
appearance of H), and vapor phase diffusion in the dredged material pore
spaces (hence, the appearance of DA3, e1? and pb).
The instantaneous flux predicted by Equation 39 decreases with time as
shown in Figure 19. Decreasing flux with time is a characteristic of
Chapter 4 Contaminant Losses During Pretreatment
-------
40 -
40 60
TIME, days
Figure 19. Predicted Aroclor 1242 flux from exposed New Bedford Harbor
Superfund sediment
contaminant volatilization from soils that is often observed in controlled labo-
ratory studies (Mayer, Letey, and Farmer 1974). The total mass loss is the
area under the curve multiplied by the surface area of exposed sediment. The
area under the curve is the integral of Equation 39 with respect to time. A
number that is useful for estimating mass loss is the average flux over some
time t' given by
rt'
« = Jo n°
dt
'' dt
(40)
Simple numerical techniques can be used to perform the integrations indicated
in Equation 40. If the top microlayer depletion is neglected, the Kg term
disappears from Equation 39. For this simplification, performing the indi-
cated integrations yields the approximate solution
Chapter 4 Contaminant Losses During Pretreatment
75
-------
na = 2 na (41)
Thus, the average volatile flux over some time t is just twice the instantaneous
flux at time t. Average flux multiplied by the area of exposed sediment and
the exposure time yields the total volatile loss.
The diffusion equation on which Equation 39 is based is well established
for pesticide volatilization from soil surfaces (Hamaker 1972; Mayer, Letey,
and Farmer 1974; Thomas 1990b) and has been successfully applied to model-
ing emissions from landfarming operations (Thibodeaux and Hwang 1982) and
hazardous waste impoundments (Dupont 1986). Solutions to the diffusion
equation involving different boundary conditions than those used in deriving
Equation 39 are available (Carslaw and Jaeger 1959) and have been applied to
modeling volatilization of pesticides from soil (Thomas 1990b).
Extrapolation of models for soils to dredged material has not, however,
been verified, and there are aspects of the simple model previously discussed
that need further development. For example, the effects of water content and
water evaporation on volatilization rates are not included in Equation 39. The
effective diffusion coefficient DA3 can be estimated by
10/3
where
DAl = air diffusivity of compound
el = air-filled porosity
e = total dredged material porosity
This relationship shows that the effective diffusion coefficient is very sensitive
to changes in the water content and porosity of the dredged material. Fully
saturated dredged material exhibits a very low diffusion coefficient. The
effects of desiccation and the subsequent reduction of porosity on volatile
emissions from dredged material have not been systematically investigated.
Since porosity is an important parameter, the assumption of constant porosity
could lead to substantial errors in volatile emission estimated from exposed
dredged material.
Thibodeaux (1989) and Taylor and Glotfelty (1988) discuss the importance
of water content and evaporation of water as factors and processes affecting
volatilization. Major differences in diurnal volatilization rates have been
76
Chapter 4 Contaminant Losses During Pretreatment
-------
observed that are related to water content. Volatilization rates decrease during
the day as the soil surface dries and increase at night as soil moisture losses
during the day are replaced by subsurface soil moisture. Volatilization rates
have also been observed to increase significantly following rainfall. The
effect is probably due to competitive adsorption between water molecules and
contaminant molecules for sorption sites on soil particles.
Evaporation induces an upward movement of water that results in convec-
tive flow of the bulk pore gas. Thibodeaux (1989) presented an enhancement
factor approach to account for evaporation that simplifies coupling convective
movement of water and diffusive movement of volatile chemicals. Convective
movement of water, however, distorts diffusive gradients, and evaporation is
not a continuously steady process. Evaporation varies greatly under field
condition and may cease at high relative humidity.
Thibodeaux (1989) also recognized desiccation cracking of the dredged
material surface as a process likely to affect volatilization and suggested some
approaches to developing volatile emission models that include the effects of
desiccation cracking. Figure 20 shows the type of desiccation cracking that
takes place in fine-grain dredged material. Such cracks can encompass up to
20 percent of the volume of the surface crust that develops by evaporative
drying (Haliburton 1978).
Figure 20. Desiccation cracking of exposed dredged material
Volatile emission summary. Predictive techniques for the ponded dredged
material and the exposed sediment volatilization locales were described. The
predictive techniques, however, are based on simple models that in some cases
do not account for important factors and/or processes. Development of
Chapter 4 Contaminant Losses During Pretreatment
77
-------
predictive models that take into account water content, water evaporation, and
desiccation cracking is a critical need for estimating volatilization losses from
exposed dredged material. Laboratory and field testing is needed to build a
higher degree of confidence in the predictive capability of the available volatil-
ization models.
78
Chapter 4 Contaminant Losses During Pretreatment
-------
Losses From Confined
Disposal Facilities
Background
Confined disposal facilities1 are often used in the Great Lakes for disposal
of dredged materials that are unsuitable for open-water disposal. When con-
taminated dredged material is placed in a CDF, contaminants may be mobi-
lized and transported away from the CDF by a variety of physical, chemical,
and biological processes. Release rates vary depending on the chemical and
engineering properties of the dredged material, the method of dredging and
dredged material placement, CDF location, stage of filling, and CDF design,
operation, and management.
Pathways involving movement of large masses of water, such as CDF
effluent, have the greatest potential for moving significant quantities of con-
taminants out of CDFs (Brannon et al. 1990). Other water-related migration
pathways include ponded water seepage through permeable dikes, seepage of
leachate through permeable dikes, seepage of leachate through foundation
soils, and surface runoff. Pathways such as volatilization may also result in
movement of substantial amounts of volatile organic chemicals at certain
stages in the filling of a CDF. Internal contaminant cycling can also be
important in the long-term mass balance for CDFs (Brannon et al. 1990).
This section begins with an overview of CDF disposal technology, fol-
lowed by a review of the literature on contaminant losses from CDFs. Predic-
tive techniques for effluent, leachate, and volatile losses, major contaminant
loss pathways for pretreatment facilities and CDFs, were discussed in Contam-
inant Losses During Pretreatment. CDFs have additional contaminant loss
pathways that must be considered—losses associated with runoff and dike
seepage. Predictive techniques for runoff and dike seepage losses are dis-
cussed in this section.
1 The terms confined disposal facility, confined disposal area, confined disposal site, diked
disposal area, containment area, and diked dredged material containment area are used inter-
changeably in the literature.
79
Chapter 5 Losses From Confined Disposal Facilities
-------
Overview of Confined Disposal Facility Technology
Contaminant releases from CDFs depend on a number of factors including
CDF design, operation, and management, nature and level of contamination in
the dredged material, and the physicochemical environment of the disposal
site. Factors related to site location and CDF design, operation, and manage-
ment are discussed in this section.
CDF siting locales
CDFs can be located in three disposal environments: upland, nearshore,
and in-water (Figure 21). Upland CDFs may be formed by construction of
earthen dikes or the use of existing pits or depressions. Nearshore and
in-water CDFs may be constructed with soil, stone, or combination soil and
stone-filled dikes. There are numerous modifications of these dike design
themes such as back-filling with stone on either side of sheet piling, cellular
sheet pile construction, placement of grout-filled fabric mattresses on
rock-filled dikes, use of geotextiles in soft foundation soils, and the use of
sand blankets and/or clay cores in the dike design.
DREDGED
MATERIAL
DISPOSAL
UPLAND
IN-WATER
NEAR-SHORE
Figure 21. Three general locales for siting CDFs
80
Chapter 5 Losses From Confined Disposal Facilities
-------
CDFs are rarely cited far away from the waterway for which the CDF
serves as a dredged material disposal site. Transportation costs for remote
sites are frequently a prohibitive factor as is the difficulty of finding suitable
remote sites. For these reasons, CDFs are usually located near or in the
waterway. Upland CDFs are generally located adjacent to the waterway for
which the CDF serves as a dredged material disposal site, with one side
usually bordering the waterway. Nearshore CDFs are located along a shore-
line with three sides bounded by water. In-water CDFs are surrounded on all
sides by water.
Physicochemical conditions. Contaminant mobilization is regulated to a
large extent by physicochemical conditions (oxidation-reduction potential, pH,
and salinity) in the sediment or dredged material (Gambrell, Khalid, and
Patrick 1978). In situ sediments normally encountered in highly industrialized
ports are fine grained, anaerobic, and near neutral pH. A thin surface layer,
usually 1 cm or less thick, may be oxidized. Beneath this surface layer,
microbial activity results in a depletion of oxygen, nitrate, and oxidized forms
of iron and manganese and accumulation of ammonia nitrogen and reduced
forms of iron and manganese. When hydraulic dredging occurs, the sediment
is vigorously mixed with overlying site water. The resulting influent to a
CDF is a mixture of reduced sediment and oxic site water. Field studies
indicate that influents have little or no dissolved oxygen (Hoeppel, Myers, and
Engler 1978), probably because the high biochemical oxygen demand of
dredged material rapidly depletes the dissolved oxygen in the site water
entrained during hydraulic dredging.
Because of the oxygen demand imposed by microbial metabolism, the
settled solids in a CDF quickly revert to the anaerobic, near neutral pH condi-
tions previously existing in the in situ sediment and remain anaerobic and near
neutrality as long as the dredged material is flooded or saturated. Contami-
nants in dredged material are generally less mobile under anoxic (flooded)
conditions than under oxidized (dewatered) conditions (Peddicord 1988).
Since the physicochemical conditions in a CDF depend on site locale and
management, there are some important differences in the long-term mobility
of some chemicals in CDFs. The basic difference between physicochemical
conditions in an upland CDF and those in nearshore and in-water CDFs is the
extent of the penetration of oxic (dewatered) conditions. Disposal in an
unlined upland CDF with permeable foundation soils results in dewatering and
oxidation of the upper portion of the dredged material profile. Complete
dewatering and oxidation is rarely achieved except with sandy sediments
placed above the water table. Upland disposal of dredged material high in
sulfur (e.g., pyrites) can result in mobilization of metals in the surface crust as
the dredged material becomes oxic (dewatered) and the pH drops due to sulfur
oxidation. These conditions are not common in CDFs containing freshwater
dredged material. Since a major portion of the dredged material profile in
most CDFs remains saturated (anoxic, neutral pH), metal mobilization is
minimized and is less significant relative to the fully drained condition
(Peddicord 1988).
Chapter 5 Losses From Confined Disposal Facilities
-------
Groundwater interactions. The three CDF siting locales differ signifi-
cantly in their interaction with groundwater (Yu et al. 1978). Figure 22 is a
generalized sketch of groundwater-CDF interactions for the three CDF siting
locales shown in Figure 21. In the upland locale, the hydraulic gradient
between inside and outside of the CDF tends to drain the CDF and create oxic
conditions in a portion of the dredged material profile. The hydraulic gradient
is much smaller in the nearshore and in-water locales, so that saturated condi-
tions are more likely to persist in the dredged material profile. For upland
and nearshore sites, groundwater impacts are possible depending on site con-
ditions. For an in-water site, groundwater, except in unusual cases, is not
significantly impacted.
UNSATURATED
— \7 —
UPLAND: CDF IS SEPARATED FROM
GROUNDWATER BY VADOSE ZONE;
FLOW IS INTO FOUNDATION SOILS AND
TOWARD GROUNDWATER.
=—SATURATED —
NEARSHORE- CDF IS PARTALLY SITED
IN SATURATED ZONE; WATER TABLE IS
SEASONALLY DEPENDENT AND FLOW IS
THROUGH SITE.
V7
IN-WATER: CDF IS SITED "IN-GRADIENT";
FLOW OCCURS WHEN OUTSIDE WATER
ELEVATION CHANGES
Figure 22. Groundwater-CDF interactions
82
Chapter 5 Losses From Confined Disposal Facilities
-------
CDFs are not usually located in groundwater recharge zones. This is
because CDFs are cited along waterways, and most waterways receive some
groundwater discharge. The impacts of groundwater discharge contaminant
losses at CDFs have not been studied extensively because contaminant losses
at CDFs are primarily governed by surface hydrology (rainfall, etc.), dredged
material properties, and CDF design.
Placement methods
Dredged material is placed in CDFs hydraulically by pipeline dredge,
hopper dredge, or scow pumpout and mechanically by bucket dredges.
Hydraulic disposal operations involve pumping dredged material into the CDF
as a slurry that is 10- to 20-percent solids by weight. Solids settle (Figure 4)
and consolidate, and water is discharged through an outlet structure or perme-
able dikes or both to make room for additional dredged material. Mechanical
dredging usually involves dredging and transfer of material to a scow using a
bucket. Dredged material may then be transferred from the scow to the CDF
by hydraulic or mechanical methods. Because mechanical disposal does not
use water for conveyance, the volume of water introduced into a CDF that
must later be discharged is significantly reduced when mechanical dredging
and disposal methods are used compared with hydraulic dredging and
disposal.
Design and operation
CDFs are built by raising dikes around a prescribed area and are designed
to retain dredged material solids while allowing the carrier water and/or water
initially present in the CDF to be released as the CDF fills with solids. The
primary design objectives are as follows: (a) provide adequate storage capac-
ity to meet dredging requirements, and (b) attain the highest possible effi-
ciency in retaining solids during filling operations (Palermo, Montgomery,
and Poindexter 1978; USAGE 1987).
Solids retention. Solids retention is important because the major fraction
of the contaminants in dredged material is bound to sediment solids (Burks
and Engler 1978). During hydraulic disposal, water and solids separate in the
CDF by gravity sedimentation, and the clarified water is the effluent that
potentially impacts surface water quality. The design fundamentals for solids
retention during hydraulic filling of CDFs were developed by Montgomery
(1978) and refined by Shields et al. (1987). Verification studies of CDF
design procedures for solids retention were conducted by Averett, Palermo,
and Wade (1988). The settling characteristics of dredged material depend on
many variables and must be determined experimentally in laboratory settling
tests for each dredging project (Montgomery 1978; Palermo, Montgomery,
and Poindexter 1978; Palermo 1986; USAGE 1987). Based on the settling
characteristics determined in laboratory tests, the residence time required for
83
Chapter 5 Losses From Confined Disposal Facilities
-------
clarification to a target effluent suspended solids concentration can be deter-
mined. This information is used to size CDFs.
Water release. Release of the water that must be discharged during filling
operations is accomplished in three basic ways. Effluent may be released
through an outlet structure(s), pervious dikes, or both. There are many ways
that these basic methods for water release are implemented, some simple and
some complicated. Outlet structures include simple overflow weirs, sand-
filled weirs, and multimedia filter cells. Pervious dikes are rock-filled
structures that can be built with sand blankets, sheet pile crowns, and other
modifications designed to control flow and/or quality of the water released.
Outlet structures and pervious dikes are not mutually exclusive, that is, a CDF
can be designed to release water through pervious dikes for a period of time,
typically until the dikes clog with dredged material solids. After that, water is
released through an outlet structure.
Literature on Effluent Losses During Hydraulic
Disposal
As previously discussed, influent and effluent flows are approximately
equal during hydraulic disposal in most CDFs. During active disposal opera-
tions at upland, nearshore, and in-water CDFs, effluent is probably the most
significant pathway through which contaminant losses occur. Assuming
inflow equals outflow and losses associated with pathways other than effluent
are negligible, the containment efficiency equation is
CEFEFF = CINFJOT " CEFF-TOT (43)
CINF,TOT
where
CEFEFF = containment efficiency based on effluent pathway only
CINF.TOT — tota' concentration of contaminants in influent, mg/l
CEFF TOT = total concentration of contaminants in effluent, mg/f
Equation 43 has been applied in several field studies to individual contami-
nants. The data are reported as contaminant-specific removal efficiencies in
percent. This literature is reviewed below.
Hoeppel, Myers, and Engler (1978)
Influent and effluent samples from nine confined disposal sites collected
during hydraulic disposal were studied by Hoeppel, Myers, and Engler
84
Chapter 5 Losses From Confined Disposal Facilities
-------
(1978). The nine sites investigated included four on the Atlantic coast, two on
the Gulf coast, one on the Pacific coast, one in the Great Lakes, and one
inland site. Field measurements included salinity, conductivity, dissolved
oxygen, and pH. Laboratory measurements included particle size, solids,
alkalinity, combined nitrogen (organic nitrogen, ammonia nitrogen, nitrate
nitrogen, and nitrite nitrogen), total and ortho-phosphate phosphorous, total
and inorganic carbon, selected pesticides (DDT, DDE, DDD, dieldrin, aldrin,
lindane, heptachlor, heptachlor epoxide, and chlordane), PCBs, oil and
grease, sulfides, major ions (calcium, magnesium, potassium, sodium, chlo-
ride, and sulfate), and trace metals (iron, manganese, zinc, copper, cadmium,
lead, nickel, chromium, mercury, arsenic, vanadium, selenium, and titanium).
This study showed that most chemical constituents in dredged material were
associated with the solids fraction, and the efficiency of contaminant contain-
ment during filling operations was directly related to the efficiency of solids
retention.
Application of Equation 43 to influent and effluent data for eight of the
nine sites is summarized in Figure 23. Reduction in total metal concentrations
for iron, zinc, cadmium, copper, nickel, arsenic, vanadium, and lead closely
followed total solids removal (96 percent). The metals that showed average
retention efficiencies of less than 90 percent included titanium (89 percent),
manganese (88 percent), potassium (78 percent), and mercury (46 percent).
Most total nutrient concentrations (total organic carbon, organic nitrogen,
and total phosphorus) showed retention efficiencies approximating total solids
removal (96 percent). Total ammonia-nitrogen removal was only 57 percent.
Oil and grease, most pesticides, and PCBs showed very efficient removal
when adequate solids retention was maintained. Almost all of the oil and
grease, pesticide, and PCB was associated with solids in both the influent and
effluent samples. Although oil and grease were efficiently removed during
dredged material containment, sediments with high contents of petroleum
residues seemed to settle more slowly, often forming an oil-water-sediment
layer near the bottom of ponded areas in the CDF.
Luetal. (1978)
Lu et al. (1978) carried out studies similar to those conducted by Hoeppel,
Myers, and Engler (1978) at two sites, one in Mobile, AL (Pinto Island), and
one in Detroit, MI (Grassy Island). This study placed major emphasis on size
fractionation of influent and effluent suspended paniculate matter. The results
showed that most trace metals, oil and grease, chlorinated pesticides, and
PCBs were almost totally associated with settleable solids (> 8 /^m) in influent
and effluent samples. A significant fraction of total calcium, magnesium,
sodium, potassium, ammonia-nitrogen, total carbon, and organic carbon was
associated with the dissolved phase (<0.05 /xm). Containment efficiencies for
these parameters were low relative to the solids retention efficiency.
85
Chapter 5 Losses From Confined Disposal Facilities
-------
% DECREASE % INCREASE
1 00 90 80 70 60 50 40 30 20 1 0 0 1 0 20 30 40 50 60 70 80 90 1 00
I I I I I I I I I I I I I I I I I I I
•••^•^^ ORGANIC -C, <0.45 p,m
CALCIUM. <045 |Xm ••••§
•• MAGNESIUM, <0.45 (0. m
POTASSIUM, <045 M-m •
•• SODIUM, <0.45 Urn
••• SODIUM, TOTAL
ZINC, <045 (im «^H
mtmmmmmm CADMIUM, <0 45 \Lm
COPPER. <0.45 Urn "•"
••••• NICKEL, <0.45 \i m
LEAD, <045 Hm
MERCURY, <045 \lm
CHROMIUM, <045 M-m ^^"
^ TITANIUM, <045 ^m
mmm^ VANADIUM. <0 45 \lm
•i CHLORIDE
• EXCHANGEABLE AMMONIUM - N + <0 45 ^m AMMONIUM - N
Figure 23. Contaminant containment efficiencies for eight CDFs (Hoeppel, Myers, and
Engler 1978 as cited by Palermo 1988)
86
Chapter 5 Losses From Confined Disposal Facilities
-------
The Grassy Island CDF is located in the Detroit River and discharges to
the Detroit River. Sampling was conducted during hopper dredging and
disposal of material from the Rouge River in Detroit, MI. Retention efficien-
cies for most trace metals, oil and grease, chlorinated pesticides, and PCBs
were very close to the total solids retention (99.7 percent) at the Grassy Island
CDF. Parameters with retention efficiencies less than 90 percent included
ammonia nitrogen (83 percent), total organic carbon (62 percent), potassium
(61 percent), total carbon (55 percent), calcium (44 percent), and magnesium
(10 percent).
The Pinto Island CDF is located in Mobile Bay, Alabama. Sampling was
conducted during hydraulic dredging and direct pipeline disposal. The reten-
tion efficiencies at the Pinto Island CDF were generally lower than those at
the Grassy Island CDF for trace metals (cadmium, 18 percent; copper, 52 per-
cent; mercury, 35 percent; nickel, 67 percent; lead, 35 percent; selenium,
39 percent; and zinc, 35 percent). Retention efficiencies for organics were
much better than for metals at Pinto Island. PCB retention efficiencies for
Aroclors 1242, 1254, and 1260 were 96, 97, and 99 percent, respectively.
Palermo (1988)
Palermo (1988) evaluated the predictive capability of the modified elutriate
and companion settling tests for effluent quality during hydraulic filling of
CDFs. Field data from five sites, four on the Atlantic coast and one on the
Gulf coast, were compared with predictions made on the basis of laboratory
data. Average containment efficiencies (for all sites) for most contaminants
were very close to the total solids retention (99.91 percent). The average
containment efficiency for metals for the five sites was 98.56. Results for
nutrients were generally similar to those for metals at most sites. PCBs were
measured at only one site, and the containment efficiency for PCBs at this site
was 99 percent.
For all five sites, the laboratory tests adequately predicted the dissolved
concentration of contaminants and the contaminant fractions of the total sus-
pended solids in the effluent. The predictions were within a factor of 1.5 of
the field data for a total of 64 of the 84 parameters measured. The modified
elutriate test was also a generally conservative predictor, that is, predictions of
effluent contaminant concentrations were generally higher than the measured
field results.
Palermo (1988) obtained detailed statistical data on the predictive capability
of the modified elutriate and companion settling tests for sites studied.
Results for both the laboratory predictions and the field data are shown in
Figures 24 and 25. In most cases, the mean of the modified elutriate was
within the standard deviation for the field data. These data provide the scien-
tific basis for recommending the modified elutriate test and companion settling
tests as the predictive techniques for estimating contaminant losses associated
87
Chapter 5 Losses From Confined Disposal Facilities
-------
LE HARBOR
m
O
1 SAVANNAH HARBOR
| NORFOLK HARBOR
PARAMETERS
TOTAL ORGANIC
CARBON
AMMONIA
NITROGEN
COPPER
IRON
MANGANESE
COPPER
IRON
LEAD
NICKEL
ZINC
TOTAL ORGANIC
CARBON
CADMIUM
CHROMIUM
COPPER
IRON
LEAD
MANGANESE
ZINC
UNITS
X10 *
mg/l
X101
mg/l
X10'1
mg/l
x10°
mg/1
X101
mg/l
X10'1
mg/l
xioo
mg/l
X10'3
mg/l
X10-2
mg/l
X10 -2
mg/l
X101
mg/l
XNT2
mg/l
1 234 5678 9
I I | «l | I I I I I I
w
n
r
I • I
I • 1
h- •— I
•
•I
•* 1
I — • — I
\—m-\
I .... m |
i,_ , • _i
i A. i
WH
1 • 1
1 • 1
M
H«H
r*H
X10'1 |-»-|
mg/l jgj
X10-2
mg/l
X10°
mg/l
X10 -1
mg/l
1- • i
H-H ' '
h^H
i ^ i
X10° I m
mg/l ||
X1CT2
mg/l
1 A , ,_]
f-»H
LEGEND
(— • — | MODIFIED ELUTRIATE DATE
| M 1 FIELD DATA
BARS INDICATE STANDARD DEVIATION
Figure 24. Means and standard deviations of predicted and observed effluent quality at
Mobile Harbor, Savannah Harbor, and Norfolk Harbor CDFs
88
Chapter 5 Losses From Confined Disposal Facilities
-------
BLACK ROCK HARBOR
1 HART MILLER ISLAND
PARAMETERS
TOTAL
PHOSPHORUS
AMMONIA
NITROGEN
TOTAL ORGANIC
CARBON
CHROMIUM
COPPER
IRON
LEAD
MANGANESE
NICKEL
TOTAL
PCB
CADMIUM
CHROMIUM
COPPER
LEAD
SELENIUM
BARIUM
IRON
UNITS
mg/l
mg/l
x10
mg/l
X10
mg/l
X10"3
mg/l
mg/l
X10
mg/l
mg/l
X10'3
mg/l
xicr3
mg/l
X10'2
mg/l
X10'3
mg/I
X1(T3
mg/l
X10'2
mg/l
XlO'3
mg/I
X10-2
mg/I
x10'2
mg/l
x10°
RELATIVE VALUE
12345678 9
1 , U | 1 1 1 1 1 1
1 • 1
1 •<-<•!
In ... •"
1 •
1 — 1
1 • |
1 a 1
1 M 1
W
• *
r^l
1 • 1
h*H
L A.. , 1
1 •<• (
1 _^ 1
• 1
I ^ |
I-W
^^
III
W«
1 "• 1
*~^ 1 • 1
I m 1 '
1 • 1 l_ m 1
r — • 1
1 • 1
*
LEGEND
|— • 1 MODIFIED ELUTRIATE DATA
1 • — 1 FIELD DATA
BARS INDICATE STANDARD DEVIATION
Figure 25. Means and standard deviations of predicted and observed effluent quality at
Black Rock Harbor and Hart Miller Island CDFs
Chapter 5 Losses From Confined Disposal-Facilities
89
-------
with effluent. The data, however, are primarily nutrients, dissolved oxygen,
pH, organic carbon, and metals concentrations in effluent. The dissolved and
total organic carbon estimates provided by the laboratory tests were in good
agreement with the field data. The modified elutriate and companion settling
tests should, therefore, be a good predictor of dissolved and total organic
chemical concentrations in effluent. Sediment from one site, Black Rock
Harbor, Connecticut, contained high enough concentrations of PCBs
(14.3 mg/kg total PCB) for PCBs to be found in the effluent during disposal
operations. The mean total PCB concentration in effluent from the Black
Rock Harbor CDF was 0.0099 mg/f, versus a predicted value of 0.013 mg/f.
Thackston and Palermo (1990)
Thackston and Palermo (1990) applied the modified elutriate and compan-
ion settling tests to prediction of effluent quality from a CDF for the Houston
Ship Channel, Texas, during hydraulic filling. This study was designed to fill
data gaps on freshwater sediments and organic contaminants. Additional
information on effluent quality during disposal of a freshwater sediment was
obtained, but the organic chemical contamination of the sediment was too low
to obtain information on the predictive capability of the modified elutriate and
companion settling test for organic contaminants in CDF effluent. In this
study, the mean ratios of predicted to observed effluent nutrient and metals
concentrations was near 1.0, and the range in predicted total to observed total
effluent contaminant concentrations was 0.2 to 2.6. Total ammonia-nitrogen
concentration was underpredicted (ratio of predicted to observed = 0.2), and
total chromium was overpredicted (ratio of predicted to observed = 2.6). On
balance, the data set obtained again showed that the modified elutriate and
companion settling tests comprise a useful and reasonably accurate predictive
technique.
Thackston and Palermo (1992)
Additional verification work on PCBs was conducted by Thackston and
Palermo (1992) at the New Bedford Harbor Superfund Demonstration CDF,
New Bedford, MA. The PCB concentrations in sediments from the site
ranged from a few milligrams per kilogram to over a gram per kilogram
(Averett 1988). A demonstration-scale CDF for hydraulic dredging and dis-
posal of 1,680 m3 of contaminated sediment was constructed as part of a pilot
study of dredging and disposal alternatives. The total PCB concentration in
the composite sample used for modified elutriate testing was 2.2 g/kg. The
predicted value for dissolved PCB (0.0075 mg/f) was very close to the
observed mean value for dissolved PCB in the CDF effluent (0.0045 mg/f )•
90
Chapter 5 Losses From Confined Disposal Facilities
-------
Myers (1991)
Myers (1991) measured PCB congener concentrations in influent and pond
water in the Saginaw CDF, Saginaw, MI. Sampling was conducted during
hopper dredging and disposal of material from the Saginaw River near
Saginaw, MI. The perimeter dike at the Saginaw CDF was designed to be
permeable. Effluent monitoring was not practical because the discharge
through permeable dikes is diffuse and is quickly diluted to background con-
centrations. Based on PCB congener concentrations in pond water collected
on the inside face of the perimeter dike, the containment efficiency of the
Saginaw CDF for PCBs was 99.82 percent. This estimate neglects filtration
and sorption in the dike and assumes that the dike is transparent to both dis-
solved and paniculate PCB.
Myers (1991) also compared PCB congener concentrations in the modified
elutriate test with observed pond water PCB congener concentrations during
disposal operations. The results of this study were consistent with the verifi-
cation studies of Palermo (1988) and Thackston and Palermo (1992), which
involved sediments with higher contamination levels and used total PCB as the
model parameter. Of the 60 PCB congeners analyzed, 16 were detected in the
unfiltered modified elutriates, compared with 13 detected in unfiltered pond
water samples. The predicted total concentrations for 4,4'-dichlorobiphenyl
and 2,2',5,5'-tetrachlorobiphenyl, the two most abundant PCB congeners in
the dredged material influent, were 0.02 and 0.07 ng/f, respectively, com-
pared with observed concentrations in the CDF pond water of about 0.05 and
0.003 fj.g/1 for 4,4'-dichlorobiphenyl and 2,2',5,5'-tetrachlorobiphenyl,
respectively. Dissolved PCB congener concentrations were generally below or
just above the chemical analytical detection limit (0.01 ng/t) in both the modi-
fied elutriate test and CDF pond water.
Krizek, Gallagher, and Karadi (1976)
Krizek, Gallagher, and Karadi (1976) studied influent and effluent samples
collected during hydraulic filling of the Perm 7 CDF in Toledo, OH. The
experimental design was similar to that used by Hoeppel, Myers, and Engler
(1978) and Lu et al. (1978) in that numerous influent and effluent samples
were collected. Samples were analyzed for metals, nutrients, chemical oxygen
demand (COD), and biochemical oxygen demand (BOD).
Containment efficiencies for most parameters were very close to the total
solids retention (99.7 percent). Average retention efficiencies were as fol-
lows: iron (99.8 percent), COD (99.1 percent), potassium (98.8 percent),
total phosphate (98.7), BOD (98.4 percent), calcium (97.5 percent), manga-
nese (96.7 percent), zinc (95.9 percent), sodium (87.5 percent), cadmium
(63.5 percent), copper (45.0 percent), and lead (45.0 percent). Effluent
nitrate-nitrogen showed a 10-fold increase over influent nitrate-nitrogen. The
authors attributed this increase to nitrification of nitrogenous compounds in the
CDF.
91
Chapter 5 Losses From Confined Disposal Facilities
-------
MacKnight and MacLellan (1984)
MacKnight and MacLellan (1984) described disposal of PCB-contaminated
sediment at Petit-de-Grat, Nova Scotia, Canada. Sediment containing 2- to
25-mg/kg PCB was hydraulically dredged and disposed in a CDF. A cationic
polymer was used to improve solids removal in the CDF. Samples collected
and analyzed for suspended solids and PCBs showed that the effluent met
effluent water quality guidelines of less than 300-mg/f suspended solids and
less than 0.05-^g/f PCBs. The authors concluded that hydraulic disposal in a
CDF is an economically and environmentally acceptable method of disposal.
Khan and Gross! (1984)
Khan and Grossi (1984) presented results from a single round of effluent
quality tests conducted during hydraulic disposal of contaminated sediment
from Hamilton Harbor, Ontario, Canada. During this disposal operation,
dredged material was pumped into a primary sedimentation cell. The superna-
tant from the primary cell traveled through three more cells before discharge
to Hamilton Harbor. The results showed solids retention of 98.5 percent and
effluent water quality comparable with ambient water conditions outside the
CDF.
Effluent Losses During Mechanical Disposal
Predictive techniques for effluent quality during mechanical disposal of
dredged material are currently unavailable. Mechanical placement of dredged
material in a CDF differs from hydraulic placement, not only in the way
placement is accomplished, but also in the way dredged material behaves once
it is in the CDF. Mechanically dredged and disposed sediments have a mark-
edly different character from hydraulically dredged sediment due to the fact
that they have not been slurried with water as part of the dredging process.
Since placement is at a much higher solids concentration, there is less efflu-
ent. In many instances, fine-grain mechanically dredged sediments have a
paste-like cohesive character. In the mechanical placement process, dredged
material primarily stays where it is initially placed, and only a very small
proportion of the solids are actually released to water that may have been in
the CDF prior to filling operations. It is therefore inappropriate to use a test
like the modified-elutriate test, which involves slurrying sediment and water to
estimate contaminant release during mechanical disposal.
Dredged material mechanically placed in upland CDFs should generate
little to no effluent for discharge. Mechanical placement of dredged material
in nearshore and in-water CDFs will displace the water initially present as
filling proceeds. Because mechanical disposal rates are much lower than
hydraulic disposal rates and most of the material stays where it is initially
placed, only weak currents from placement point to discharge point are
92
Chapter 5 Losses From Confined Disposal Facilities
-------
generated. The advective velocity imparted by mechanical disposal operations
is essentially negligible when the discharge point is a long distance from the
placement point.
There are three primary mechanisms by which pollutants in mechanically
placed dredged material are released to ponded water in nearshore and
in-water CDFs. These are diffusion, release of pore water by consolidation,
and resuspension of fine solids. Probably the most important process is wind-
induced resuspension. Wind-induced currents resuspend sediment solids and
disperse contaminants released by diffusion and pore water released by consol-
idation. Without wind-induced currents, migration of contaminant to an efflu-
ent discharge point is extremely slow.
Jones and Lee (1978) proposed development of a plop test for estimating
contaminant release during mechanical disposal of dredged material. A plop
test has never been developed, and the amount of testing conducted by Jones
and Lee (1978) was limited. Some Corps of Engineers Districts have esti-
mated effluent quality during mechanical disposal in an in-water CDF as
dilution of pore water by ponded water in the CDF. This method is maybe
better than no method at all; but since resuspension is not accounted for, this
method underestimates pollutant releases.
Basically, there are two approaches to developing a predictive test, and the
approach taken significantly affects test design and the basis for extrapolating
laboratory data to the field. One approach is strictly empirical. It uses statis-
tical analysis to establish correlation between laboratory and field data. The
other approach is deterministic. In the deterministic approach, a mathematical
model is derived from the physical-chemical laws that govern important pro-
cesses. The mathematical model will require some parameter estimation and
is therefore not purely deterministic. Most predictive techniques embody a
combination of approaches with one being the primary basis for experimental
design.
In the case of an empirical approach, a laboratory test should simulate the
placement process, release of pollutants, and transport to the discharge point.
In general, a laboratory test can never fully simulate all the minutia of field
phenomena. With sufficient laboratory and field data, however, correlation
functions can be developed that provide a basis for prediction. The cost of
obtaining enough reliable data is a disadvantage of the empirical approach.
Another disadvantage is that unless the laboratory test simulates important
processes, the correlation functions may be too statistically weak to be of
practical value.
The deterministic approach involves describing the pollutant release-
transport-discharge process mathematically, assigning coefficients or variables
to each part of the overall process, and estimating the magnitude of each
coefficient or variable. The entire process may never be experimentally simu-
lated as a whole, but, instead, each step is simulated or analyzed separately;
the steps are then combined logically and/or mathematically. In order to
93
Chapter 5 Losses From Confined Disposal Facilities
-------
successfully implement this type of approach, the overall process must be
understood well enough so that it can be broken down into a small number of
steps that can be isolated and measured in a laboratory test. The feasibility of
this type of approach is enhanced if the overall process depends primarily on
one or two mechanisms such that other mechanisms are unimportant or they
have a range of variability so low that they can be assumed to be constant.
Development of the procedures for predicting effluent quality during
hydraulic disposal previously discussed is an example of a successful combina-
tion of empirical and deterministic approaches. Several factors contributed to
the successful development of these procedures. First, hydraulic dredging
tends to homogenize variations in sediment chemical and physical properties
so that the use of average values is consistent with the physics of the process.
In addition, many of the variables affecting contaminant release during
hydraulic filling are understood because of considerable experience with
hydraulic dredging. Flocculation and sedimentation have been studied for
many years so that there was a large knowledge base from which to initiate
test development. Further, the time scale of the flocculation-sedimentation
process is large relative to the time required for many individual chemical or
physical reactions so that minor errors in variable estimation are not critical.
The above discussion describes technical aspects of developing a predictive
technique for effluent quality during mechanical disposal. The problem is not
sufficiently understood to determine which of the two approaches discussed
should be recommended. Since mechanical dredging and disposal is an alter-
native that is sometimes selected for contaminated sediments, development of
a predictive technique for effluent quality during mechanical disposal in near-
shore and in-water CDFs is needed to fully evaluate this alternative.
Seepage Through Permeable Dikes:
Nearshore and In-Water CDFs
Pond water seepage through dikes
Some nearshore and in-water CDFs use permeable dikes to release the
carrier water used in hydraulic dredging. Figure 26 is a typical cross section
of the perimeter dike at the Saginaw CDF, Saginaw, MI. Dredged material
solids clog permeable dikes as CDFs fill so that an outlet structure(s) is
usually necessary for release of carrier water in the latter stages of filling.
During disposal operations, the flow through the dike is the volumetric
influent flow if the influent flow is continuous. In between disposal opera-
tions and when influent flow is not continuous, there is a potential for lake
water to move through the dike into the CDF and then back out again as
lakeside water levels fluctuate. The direction of flow depends on water eleva-
tions inside and outside the CDF. Flow through the dike can be estimated
using Dupuit's equation, Equation 44.
94
Chapter 5 Losses From Confined Disposal Facilities
-------
TOP OF DIKE
S'-O"
NEATLINE
COVER STONE
UNDER LAYER STONE
LW.D. 0.0'
PREPARED
LIMESTONE v
PLASTIC FILTER CLOTH
PLASTIC FILTER CLOTH
I
r-T RIPRAP STONE
(f-4" TO r-8")
BOTTOM ELEV. VARIES
• MATTRESS STONE
• EXISTING LAKE BOTTOM (ELEVATION VARIES)
TYPICAL DIKE DETAIL
Figure 26. Cross section of perimeter dike at Saginaw CDF
q =
K (hf -
(44)
where
q = discharge per unit length of dike, m2/sec
K = hydraulic conductivity of dike, m/sec
A! = pond water elevation above base of dike, m
H2 = outside water elevation above base of dike, m
L = horizontal distance separating surface of pond and surface of
outside water body, m
A definition sketch for Dupuit's equation is given in Figure 27. The assump-
tions on which Equation 44 is based are discussed by Harr (1962).
To use Dupuit's equation, water level fluctuations outside the CDF are
needed. These data are not easily obtained for preproject analysis of contami-
nant losses. There may be several ways of dealing with this problem. Two
are briefly mentioned as follows: use historical data or develop a synthetic
water level generator based on historical data. In either case, the time scale
for the lakeside water elevations must properly represent the dispersion effect
that changing water elevations in the lake have on contaminant movement
from the pond water through the dike. There are, however, no data on con-
taminant movement through permeable dikes due to fluctuating lake levels that
can be used as guidance. Engineering judgment in the selection and use of
water level data is therefore required.
Chapter 5 Losses From Confined Disposal Facilities
95
-------
CDF POND
S«»^
Figure 27. Definition sketch for application of Dupuit's equation
To estimate the mass of contaminant released, predicted flows must be
coupled with estimates of pond water contaminant concentrations. If pond
water contaminant concentrations are assumed to be constant and equal to the
concentrations predicted by the modified elutriate test, then the mass of con-
taminant released is total flow out of the CDF times the contaminant concen-
trations predicted by the modified elutriate test. This type of estimate is
probably a crude overestimate of contaminant losses. Simple techniques for
predicting the time dependency of pond water contaminant concentrations are
not available.
Prediction of contaminant losses due to changing lake levels can also be
developed by modeling the dispersive effect of water moving in and out of the
CDF as a diffusion process. This approach is well established in estuary
modeling where the overall flow is out to sea but there is substantial mixing
by tidal effects. Martin, Ambrose, and McCutcheon (1988) incorporated
algorithms into the Water Analysis Simulation Program, Version 4 (WASP4)
that parameterized the dispersive effects of changing water level elevations in
a dispersion coefficient. This model does not require time-dependent lake
elevations as input and can simulate some of the processes affecting contami-
nant concentrations in pond water. There are, however, no data on the dis-
persion effects of fluctuating water levels in permeable dike CDFs on which to
base estimation of dispersion coefficients, nor are data available on processes
affecting contaminant concentrations in pond water. Application of WASP4
and similar models, therefore, requires engineering judgment in the selection
of dispersion and other transport process coefficients.
Leachate seepage through dikes
As previously mentioned, some nearshore and in-water CDFs use perme-
able dikes to release the conveyance water used in hydraulic dredging. Once
the CDF is filled above the high-water datum, exchange of water between the
CDF and the outside water body is restricted by the dredged material that fills
96
Chapter 5 Losses From Confined Disposal Facilities
-------
the voids in the inside face of permeable dikes. The HELP model previously
discussed provides an estimate of the total seepage likely to occur but does not
indicate the fraction that seeps through dikes. HELP model leachate flow
predictions have been interpreted to represent the total leachate released
through all boundaries of the CDF without implying that leachate only flows
vertically (Averett et al. 1988). However, when flow is two- or three-
dimensional, caution must be exercised when using a one-dimensional tool
such as the HELP model to estimate flow. If a barrier soil with a hydraulic
conductivity lower than that of the dredged material is constructed, leachate
flow into the foundation soils can be reduced. However, flow through the
dikes may be increased, depending on the hydraulic conductivity of the dikes.
The HELP model is an appropriate tool for predicting total leachate flow and
evaluating the effectiveness of a barrier soil to reduce flow into the foundation
soils, but it is not designed to provide information on potential flow through
the dikes.
Unconfined-saturated flow groundwater models are available that could be
used to model dike seepage. Such models require substantial site-specific data
on local hydrogeology. Although not described in this report, two- and possi-
bly three-dimensional models may be needed to fully describe dike seepage at
CDFs. Simplified models could also be developed for comparison with HELP
model estimates. An example of the type of simple seepage models that could
be developed is described below.
At some point in time, the amount of water entering the dredged material
as percolation and the amount leaving as leachate flow will tend to balance so
that a quasi-equilibrium exists. When a quasi-equilibrium exists, flow aver-
aged over an extended time scale is steady and, under certain conditions, is
parallel to the bottom of the CDF. Definition sketches for horizontal-steady
flow in upland and in-water CDFs are given in Figure 28. For homogeneous,
isotropic, circular CDFs, flow is radially symmetric. Radially symmetric,
steady flow in homogeneous and isotropic media is given by the following
equation (Glover 1974; McWhorter and Sunada 1977):
Q = 7T __
' ~ (45)
In
where
•K = 3.1459...
Hl = head at crown of water table mound (Figure 27), cm
H2 = head outside CDF (Figure 28), cm
97
Chapter 5 Losses From Confined Disposal Facilities
-------
WATER TABLE PLATEAU
WATER TABLE PLATEAU
r\
-T\
jfl
Figure 28. Definition sketches for horizontal-steady flow in CDFs
R2 = distance from center of CDF to dike, cm
Rl = distance from center of CDF to edge of water table crown, cm
Application of Equation 44 involves the following assumptions:
a. Isotropic and homogeneous medium.
b. Piezometric plateau in center of CDF.
c. Time invariant piezometric surface.
d. Time invariant dredged material hydraulic properties.
e. Radial symmetry with the center of symmetry coincident with the
center of the CDF.
/. Dikes with infinite permeability relative to the permeability of the
dredged material.
g. Dupuit-Forchheimer assumptions:
98
Chapter 5 Losses From Confined Disposal Facilities
-------
(1) Equipotential lines perpendicular to the bottom of the CDF.
(2) Hydraulic gradient is equal to slope of the free piezometric sur-
face and invariant with depth.
These conditions are not always met, but when they are, flow is horizontal
and modeled by Equation 45.
It is anticipated that leachate flow estimates provided by Equation 45 will
be substantially less than leachate flow estimates provided by the HELP
model. The differences are primarily due to differences in hydraulic
gradients. Horizontal hydraulic gradients as indicated in Figure 28 are
roughly a factor of 100 times lower than the vertical hydraulic gradients in the
HELP model. Because site conditions that provide horizontal-steady flow
minimize hydraulic gradients, estimates provided by Equation 45 are probably
lower bounds on leachate flow.
The analysis of horizontal-steady flow discussed above assumes an equilib-
rium between surface recharge and seepage. Such conditions are rarely estab-
lished, as there are seasonal as well as daily fluctuations in the piezometric
surface in a CDF. Equation 45 is, therefore, limited to estimation of annual
average flow in relatively old CDFs for which there is relevant site-specific
information.
The flow domain and boundary conditions at many CDFs are such that
leachate flow is not primarily vertical or steady-horizontal. Two- and three-
dimensional flow in the subsurface environment has been considered in detail
by many researchers (Harr 1962; McWhorter and Sunada 1977; Freeze and
Cherry 1979; Bear and Verruijt 1987; Strack 1989; National Research Council
1990). The literature contains many different two- and three-dimensional
numerical models of subsurface flow that could be used to analyze more com-
plicated seepage conditions in CDFs.
The main difficulty with applying these models to CDFs is that local clima-
tology and surface hydrology are not explicitly modeled. Infiltration is
usually treated as an external input requirement without accounting for the
stochastic character of rainfall events and resulting infiltration. As previously
discussed, local climatology and surface hydrology are important because the
water budget in a CDF is surface driven. Percolation to the saturated zone
and the depth of the saturated zone depend on infiltration, which depends on
the amounts of rainfall, runoff, and antecedent soil water. Since infiltration is
the long-term source of water for leachate generation, climatologic and surface
hydrologic modeling such as provided by the HELP model is a necessary
component if the analysis of leachate flow is to be complete.
A careful scientific investigation calls for the complete use of the most
up-to-date theoretical formulation and modeling tools. Preproject estimation
of contaminant losses for planning level assessments sometimes may indicate
the need for careful investigation of losses along some pathways. Losses
99
Chapter 5 Losses From Confined Disposal Facilities
-------
through permeable dikes is a contaminant loss pathway where the simple
equations previously discussed are likely to be inadequate. As an alternative,
the two-dimensional, finite element model, SEEPU, is available (Kuppusamy
1991) for estimating flow through dikes. This model has preprocessors and
postprocessors to facilitate data input and output and runs on MS-DOS based
personal computers.
Complex models are generally expected to have a greater predictive capa-
bility than simple models and increase the range of situations that can be
described. Complex models require proper input information, as obtained
from detailed field measurement. These measurements are usually quite
extensive especially if a three-dimensional model is used.
In addition, the many models available differ from one another as a result
of different objectives of the modeling effort. For this reason, a model should
not be applied unless the objectives, model structure, type of output, and
model precision are commensurate with the information needs and site condi-
tions for a particular problem. For these reasons, development of predictive
techniques for complicated flow problems in CDFs is not a search for one
correct and completely general set of equations.
Estimation of contaminant losses associated with leachate seepage through
dikes will involve coupling flow with leachate quality. Techniques for pre-
dicting leachate quality were discussed in Contaminant Losses During Pre-
treatment. These techniques are applicable to seepage from the anaerobic
zone, that is, the saturated zone. Techniques are also available for predicting
leachate quality from unsaturated, oxic dredged material crusts that develop in
CDFs during evaporative drying (Environmental Laboratory 1987; Myers and
Brannon 1988). These techniques can provide the leachate quality information
needed to estimate losses during rainfall events that produce horizontal, satu-
rated flow in the surface crust. These events must be short term in order for
the aerobic leachate quality estimates to be applicable. The procedures for
estimating leachate quality from aerobic dredged material are not discussed in
this report because techniques for predicting the companion flow needed for
contaminant loss estimation are not available.
Contaminant attenuation in permeable dikes
A parcel of water moving through a permeable dike takes a tortuous path
before finally exiting the dike. The contaminants in such a hypothetical parcel
of water are not likely to be conservatively transported. There are at least
four processes that can attenuate transport of contaminants through dikes.
These are filtration, adsorption, bioabsorption, and biodegradation. Filtration
of solids is generally recognized as the primary removal process in permeable
dikes. If dikes did not remove solids, permeable dike CDFs could not be
filled. Adsorption can remove dissolved contaminants left after solids
removal, but permeable dikes (sand and stone) have low sorption properties.
Adsorption is probably insignificant in permeable dikes.
Chapter 5 Losses Fiom Confined Disposal Facilities
-------
Bioabsorption and biodegradation are potentially significant removal pro-
cesses that have not been investigated in permeable dikes. Ponded water in
CDFs contains bacteria, protozoa, and other microscopic organisms that are
also probably present in the dikes. Because filling operations are often inter-
mittent, there is a potential for development of biofilms on dike materials.
Biofilms in dikes potentially bioabsorb (remove) and degrade (treat) dissolved
chemicals in pass-through water. Bioadsorption and biodegradation in perme-
able dikes have not been investigated. Consequently, removal and treatment
of pollutants by biofilms in dikes have been generally ignored. Models, such
as WASP4, that already have algorithms accounting for these processes need
field data on process descriptors to improve their application to CDFs.
Literature on Leachate Losses From CDFs
There have been relatively few studies of the impacts of dredged material
disposal in a CDF on groundwater and underlying soils. Some field and
laboratory work was accomplished under the DMRP, but this research was
limited in the number of sites investigated, duration of study, and number of
chemical parameters studied. Recently, research toward development of
predictive techniques for leachate quality in CDFs has been initiated under the
LEDO program. This work, which involves both theoretical and laboratory
studies, is still developmental. Some limited field data on leachate generation
in a CDF have been reported by the U.S. Army Engineer District, Buffalo.
The available information is reviewed below with emphasis on information for
the Great Lakes.
Field studies
Yu et al. (1978). Yu et al. (1978) conducted field investigations of leach-
ate impacts at four sites as follows: Sayerville, NJ; Houston, TX; Mobile,
AL; and Grand Haven, MI. At each of the four sites, dredged material and
soil samples were obtained from locations that would indicate lateral and
vertical migration of contaminants. Groundwater samples were obtained from
within the sites and directly below the sites, as well as from upgradient and
downgradient locations. Groundwater samples were collected four times in
9 months; soil and dredged material samples were collected during the first
sampling visit. Groundwater samples were filtered (0.45 fim) prior to
analysis.
The general findings of Yu et al. (1978) indicated that leachate quality is a
function of the physical and chemical nature of the dredged material, site-
specific hydrogeological patterns, and environmental conditions of the area
surrounding the site (e.g., physical and chemical nature of the adjacent soils).
The study showed degradation of groundwater quality due to dredged material
disposal in CDFs. Significant increases in chloride, potassium, sodium, cal-
cium, magnesium, total organic carbon (TOC), alkalinity, iron, and manga-
nese concentrations were measured in some downgradient groundwaters. Iron
Chapter 5 Losses From Confined Disposal Facilities
-------
and manganese appeared to be produced by localized environmental condi-
tions, and their mobility was not considered directly related to dredged
material disposal activities. Concentrations of chlorinated hydrocarbons,
cadmium, copper, mercury, lead, zinc, phosphate, and nickel in CDF leachate
were generally very low and did not appear to pose groundwater quality prob-
lems. Heavy metals were mostly in the parts-per-billion or subparts-per-
billion range. No soluble chlorinated hydrocarbons were observed in
groundwater.
Analysis of onsite dredged material and offsite soils failed to show system-
atic changes in chemical constituents. For most parameters, both increases
and decreases in values occurred in different locations as well as at different
depths. Total chlorinated hydrocarbons were higher in the dredged material
than in offsite samples. The upper soil samples generally contained higher
concentrations of chlorinated hydrocarbons than the samples obtained a few
feet below. There was no evidence of chlorinated hydrocarbon migration
from CDFs.
The Grand Haven CDF studied by Yu et al. (1978) is the same CDF stud-
ied by Hoeppel, Myers, and Engler (1978). This CDF is located on the bank
of the Grand River, Michigan. Prior to filling, the site was used for disposal
of construction debris. Onsite and offsite borings indicated that the foundation
consists of fine to coarse sands contiguous to a depth of 6.1 m where a dense
clay stratum (tens of meters thick) is encountered. Groundwater levels
measured on four different dates at nine locations in and around the CDF indi-
cated a gentle groundwater gradient through the site and toward the Grand
River. Figure 29 shows groundwater contours and directions of flow for a
typical survey. As shown in Figure 29, groundwater flows through the CDF
from east to southwest.
There was no evidence of chloride, sodium, calcium, phosphate, mag-
nesium, iron, manganese, mercury, lead, or zinc leaching from the Grand
Haven CDF. Alkalinity was higher in the dredged material leachate than in
downgradient samples. Comparison of samples collected beneath the site with
upgradient samples showed that the average values were in decreasing order
as follows: undersite, downgradient, and upgradient. This concentration
gradient indicates an alkalinity plume beneath the CDF that is diluted as it
moved downgradient. TOC was highly correlated with alkalinity in this
study. At the Grand Haven site, TOC showed a concentration gradient simi-
lar to that for alkalinity, indicating leaching of TOC along with alkalinity from
the CDF.
Downgradient cadmium concentrations were higher than in the dredged
material leachate. The difference, 0.0006 mg/t, was statistically significant,
but such a small difference is probably not environmentally significant. The
mean upgradient and downgradient cadmium concentrations were the same
(0.0014 mg/f), indicating no impact by the CDF. Copper was higher in the
dredged material leachate (0.019 mg/f) than in the downgradient samples
(0.010 mg/f). Upgradient copper concentrations were similar in copper
102
Chapter 5 Losses From Confined Disposal Facilities
-------
OH. TANKS
DRAINAGE
DITCH
VERPLANK
COAL AND DOCK CO.
95' — GROUNDWATER CONTOURS
DIRECTION OF FLOW
300 FT
Figure 29. Water level contours at Grand Haven CDF (from Yu et al. 1978}
Chapter 5 Losses From Confined Disposal Facilities
103
-------
concentration to samples collected beneath the site, suggesting that copper was
not leaching from the CDF.
A concentration gradient for nickel from the dredged material leachate to
downgradient monitoring wells was found, indicating a potential for migra-
tion. Average nickel concentrations were higher in the dredged material
leachate (0.127 mg/f) than in the groundwater beneath the site (0.065 mg/t)
and the downgradient groundwater (0.027 mg/f). However, because the
average nickel concentrations in the upgradient wells (0.170 mg/l) were
higher than in the CDF leachate, the CDF may not be the primary source for
nickel beneath the site and downgradient.
Leonard (1988). Leonard (1988) reported significant heavy metal and
organic contamination in pore water in dredged material in the Times Beach
CDF, Buffalo, NY. The Times Beach CDF is located on Lake Erie and was
used for confined disposal of contaminated dredged material from the Buffalo
River, the Buffalo Harbor, and the Black Rock Channel from 1972 to 1976.
The site is underlain by fine sands, glacial till, and limestone. Upgradient
monitoring wells showed evidence of arsenic, cadmium, and lead contamina-
tion. Groundwater beneath the site showed little evidence of contamination.
Krizek, Gallagher, and Karadi (1976). The field investigation conducted
by Krizek, Gallagher, and Karadi (1976), previously discussed in the section
on effluent losses, included some limited groundwater sampling within the
vicinity of the Perm 7 CDF in Toledo, OH. The quality of the groundwater
was found to be slightly worse than either the river water or the CDF efflu-
ent. Seepage from the CDF into the underlying soil was thought to be small
due to the low permeability of the dredged material and the upper strata of the
foundation soils.
Laboratory studies
Mang et al. (1978). Mang et al. (1978) investigated the generation of
leachate from 16 large plexiglass lysimeters under various environmental con-
ditions. The study used dredged material from five different locations and
two native soils from California. Various leaching solutions were used includ-
ing distilled water (rainwater leach), distilled water acidified to pH 4.5 with
sulfur dioxide (acid rainfall leach), hard water buffered with bicarbonate
(alkaline groundwater leach), and leachate obtained from a solid waste land-
fill. Parameters analyzed included major cations and anions, trace metals,
PCBs, chlorinated pesticides, nutrients, and gross physicochemical parameters
(Eh, pH, alkalinity, and conductivity).
The results showed that no single mechanism governs contaminant leaching
from dredged material. During leaching some parameters increased (Eh, pH,
TOC, alkalinity, and manganese), some remained relatively constant (phos-
phorus and magnesium), some decreased (organic and ammonia nitrogen,
copper, calcium, sodium, and potassium), some parameters were highly
104
Chapter 5 Losses From Confined Disposal Facilities
-------
variable (cadmium and zinc), and some were consistently below detection
limits (PCBs and chlorinated pesticides). This work showed that alkalinity,
iron, manganese, zinc, and lead posed the greatest potential for dredged
material disposal in a CDF to adversely impact groundwaters.
Soils placed beneath the dredged material tended to regulate pH, TOC, and
alkalinity and serve as a source for iron, manganese, calcium, potassium,
nitrate-nitrogen, and total Kjeldhal nitrogen. Adsorption onto soil solids
seemed to be an important mechanism for attenuation of ammonia nitrogen,
cadmium, copper, mercury, and lead.
Environmental Laboratory (1987). In this comprehensive study of dredg-
ing and disposal alternatives for PCB-contaminated sediment in Indiana Har-
bor, Indiana, batch and column leaching studies were conducted. The results
showed that the metal and organic contaminants in Indiana Harbor sediment
were tightly bound to the sediment solids. Less than 1 percent of the bulk
metal concentrations were leachable in sequential batch leach tests. The over-
all batch equilibrium distribution coefficients for PCBs was very high,
256,000 £/kg, indicating a low potential for leaching. Integration of batch
and column test data using a mass transport equation showed that contaminant
interphase transfer could be modeled using classical partitioning theory. Total
PCB concentrations in leachate from a CDF containing Indiana Harbor
dredged material were predicted to not exceed 0.0005 mg/f. Metals were
predicted to be near detection limits in leachate from CDFs filled with Indiana
Harbor dredged material. The results also showed significant mobilization of
metals in sediment that had been treated to simulate physicochemical condi-
tions in the oxic crust that develops during evaporative drying.
Myers and Brannon (1988). Myers and Brannon (1988) conducted batch
and column leach tests on New Bedford Harbor Superfund Site sediment.
Desorption of PCBs and metals did not follow classical partitioning theory.
Anaerobic desorption isotherms showed nonconstant partitioning for PCBs and
metal during sequential leaching. Nonconstant partitioning in this sediment
was due to salinity dependent release of sediment organic carbon (Brannon et
al. 1991). Observed and predicted column elution curves qualitatively agreed,
but quantitative agreement was not good. Predictions based on batch tests
generally overpredicted observed column leachate contaminant concentrations.
Salinity-dependent nonconstant partitioning is not expected to occur in the
freshwater sediments and dredged materials in the Great Lakes.
Palermo et al. (1989). Palermo et al. (1989) conducted batch and column
leach tests on sediment from Everett, WA. The contaminant levels in this
sediment were low relative to those in Indiana Harbor and New Bedford Har-
bor sediments. Many contaminants leached in amounts below or near the
chemical analytical detection limits. Results for contaminants that leached in
amounts that could be reliably quantified were similar to those from New Bed-
ford Harbor sediment. Salinity-dependent nonconstant partitioning was again
observed.
105
Chapter 5 Losses From Confined Disposal Facilities
-------
Literature on Volatile Losses From CDFs
There are very few field data on volatile emissions from CDFs in the
literature. Semmler (1990) did a desktop evaluation of the relative signifi-
cance of PCB volatile losses from an upland and an in-water CDF filled with
dredged material from Indiana Harbor, Indiana. This analysis indicated that
volatile PCB losses from an upland CDF were approximately four times the
volatile PCB losses from an in-water CDF. This analysis also indicated that
volatile PCB losses from both disposal locales were three orders of magnitude
higher than the PCB losses associated with leaching and four orders of magni-
tude higher than PCB losses associated with dike seepage. Semmler (1993)
conducted field studies at a CDF in which PCB concentrations in sediment,
water, and air compartments were monitored. The field results showed that
the volatile pathway accounted for the majority of PCB mass loss from May
to October. The studies of Semmler (1990) and Semmler (1993) serve notice
that the volatile emission migration pathway could be of major significance for
PCBs and other hydrophobic organic chemicals in CDFs.
EBASCO Services Incorporated (1990) conducted an ambient air monitor-
ing program for the New Bedford Harbor Superfund pilot CDF. This study
showed some of the pitfalls of attempting to measure emission rates by ambi-
ent air monitoring. PCB concentrations in ambient air around the site before,
during, and after dredging and disposal activities were indistinguishable.
These data should not be construed to imply that PCBs were not released to
the air during dredging and disposal. Changing meteorological conditions,
specifically wind velocities, generate turbulence that transports chemicals in
all directions on a local scale. The-result is a large and confusing data set
when surface samplers are placed around a site with a large emission surface
area and significant potential for high background levels. The upgradient and
downgradient concepts applicable to groundwater and surface water monitor-
ing are difficult to apply to air monitoring on a local scale.
Literature and Predictive Techniques for Runoff
Losses
As previously discussed, when dredged material is placed in CDFs, physi-
cochemical changes associated with evaporative drying affect contaminant
mobility, including surface runoff quality. This section discusses techniques
for predicting runoff quality from dredged material. Surface runoff flow
predictions from CDFs can be obtained using the HELP model previously
discussed.
Newly dredged sediment is generally anaerobic with near neutral pH and
has high water content. During the wet, anaerobic stage, the transport of
contaminants in surface runoff is mainly through the transport of suspended
solids. As the material dries and oxidizes, the pH can decrease to sometimes
1 OR
Chapter 5 Losses From Confined Disposal Facilities
-------
as low as 4 when high concentrations of sulfides are present. During the wet,
anaerobic stage, metals tend to be bound as low solubility metal sulfides. As
the dredged material oxidizes, some of these metals may increase in solubility
and be released during storm events.
WES Rainfall Simulator-Lysimeter System
The WES Rainfall Simulator-Lysimeter System (RSLS) combines a rainfall
simulator with a lysimeter bed containing dredged material (Figure 30). With
the WES RSLS, runoff samples can be collected for analysis during simulation
of selected storm events. By allowing the material placed in the lysimeters to
age, changes in runoff quality as dredged material dries can be determined.
RAINFALL SIMULATOR
LYSIMETER UNIT2
LYSIMETER UNIT 1
^^r^ *-*=*
N VARIABLE SLOPE AND
DEPTH SOIL LYSIMETER
• RUNOFF QUANTITY AND
QUALITY MONITORING
Figure 30. Schematic of WES Rainfall Simulator-Lysimeter System (from Skogerboe et al.
1987)
The rainfall simulator is a modified version of a rotating disk rainfall simu-
lator originally developed at the University of Arizona (Morin, Goldberg, and
Seginer 1967). Until the rotating disk-type simulator was developed, rainfall
simulators were unable to simulate the kinetic energy of natural rainfall
(Morin, Cluff, and Powers 1970). The rainfall simulator used in the WES
RSLS is equipped with several important design modifications, including a
Chapter 5 Losses From Confined Disposal Facilities
107
-------
programmable slit disk opening that can instantly change rainfall intensity
(Westerdahl and Skogerboe 1982). The WES rainfall simulator has been
tested and calibrated to optimize drop-size distribution, terminal drop velocity,
and rainfall intensity distribution (Skogerboe et al. 1987).
The lysimeters used in the WES RSLS are constructed of aluminum with
surface dimensions of 4.6 by 1.2 m. Depth is adjustable in 15-cm increments
to 1.2 m, and slope can be varied from 0 to 20 percent.
Runoff quality studies using WES Rainfall Simulator-Lysimeter System
Verification studies. A series of field verification tests were conducted by
Peters, Lee, and Bates (1981) and Lee and Skogerboe (1984) that showed that
the WES RSLS could accurately simulate surface runoff from natural storm
events under a variety of conditions. The effect of plant biomass on runoff
suspended solids concentrations was a major focus of these studies.
Skogerboe et al. (1987). Skogerboe et al. (1987) evaluated surface runoff
water quality impacts from an upland dredged material disposal site at Black
Rock Harbor, Bridgeport, CT, using the WES RSLS. This work was con-
ducted as part of the U.S. Army Corps of Engineers/U.S. Environmental
Protection Agency Interagency Field Verification of Testing and Predictive
Methodologies for Dredged Material Disposal Alternative Program (Field
Verification Program (FVP)). Sediment was collected from Black Rock Har-
bor and tested at WES to predict surface runoff water quality. Similar mate-
rial was also dredged from Black Rock Harbor and disposed in an upland
disposal site. Laboratory and field results showed significant increases in the
mobilities of cadmium, copper, nickel, zinc, and manganese as the dredged
material aged. Statistical analysis of observed and predicted runoff quality
showed no significant differences. Results of this study, therefore, demon-
strated that the WES RSLS can simulate the physicochemical changes and
resulting changes in runoff quality that take place when contaminated dredged
material is placed in upland environments.
Environmental Laboratory (1987). In the comprehensive study of dredg-
ing and disposal alternatives for PCB-contaminated sediment in Indiana Har-
bor, Indiana (Environmental Laboratory 1987), the WES RSLS was used to
evaluate potential runoff water quality impacts. The results showed that dur-
ing the early, wet, anaerobic stages, contaminants were primarily bound to the
suspended solids in runoff. Filtered concentrations during this period were
low compared with unfiltered concentrations, but were still of concern when
compared with the USEPA Maximum Criteria for the Protection of Aquatic
Life. As the sediment dried, the suspended solids concentrations decreased,
thereby decreasing the unfiltered contaminant concentrations.
After the sediment dried and aged for 6 months, water quality constituents
in runoff changed. Organic contaminants were no longer a concern because
most of these compounds had been lost by volatilization and/or
1 08
Chapter 5 Losses From Confined Disposal Facilities
-------
biodegradation. No PCBs were detected in runoff from dry, oxidized sedi-
ment. Heavy metals concentrations also decreased; however, many became
more soluble. Filtered concentrations of cadmium, copper, nickel, zinc,
manganese, and lead were not significantly different from unfiltered concen-
trations, indicating that these metals were primarily present in soluble form.
Filtered concentrations of cadmium, copper, zinc, and lead were greater than
or equal to the USEPA Maximum Criteria for the Protection of Aquatic Life.
Palermo et al. (1989). In the evaluation of dredged material disposal
alternatives at Everett, WA (Palermo et al. 1989), the WES RSLS was used to
evaluate potential runoff water quality impacts. The results showed that dur-
ing the early, wet, anaerobic stages, contaminants were primarily bound to the
suspended solids in runoff. All filtered metal concentrations were signifi-
cantly less than the USEPA Maximum Criteria for the Protection of Aquatic
Life and were not considered a problem as long as the dredged material
remained wet and anaerobic. Organic contaminant concentrations were also
low, especially in filtered samples. PCBs were below the detection limits in
both unfiltered and filtered samples.
After 6 months of drying and aging, the sediment did not form the hard
crust with large cracks typical of many sediments. The material remained
light and fluffy and was highly susceptible to erosion with suspended solids in
runoff averaging 1,000 mg/£. The sediment pH also remained high. Heavy
metal concentrations in filtered samples were significantly lower than concen-
trations in unfiltered samples, indicating that the major fraction was in particu-
late form. However, filtered concentrations of some metals were high. Fil-
tered concentrations of cadmium were significantly greater than the USEPA
Maximum Criteria for the Protection of Aquatic Life, and filtered concentra-
tions of copper and zinc were not significantly different from the criteria.
Unfiltered and filtered concentrations of polynuclear aromatic hydrocarbons
(PAHs) were very low, and PCBs were below the detection limit.
In addition to providing information on runoff quality, Palermo et al.
(1989) made estimates of yearly mass release for an upland CDF. These
predictions were calculated using the Universal Soil Loss Equation
(Wischmeier, Johnson, and Cross 1971). Annual losses for cadmium, copper,
zinc, and lead were estimated to be 6.2, 2.4, 115, and 0.7 kg/ha, respectively.
The estimates involve using a soil erodibility factor obtained from the RSLS
tests and a site-specific rainfall erodibility factor in the Universal Soil Loss
Equation.
Skogerboe, Price, and Brandon (1988). Skogerboe, Price, and Brandon
(1988) conducted surface runoff tests on New Bedford Harbor Superfund Site
sediment with PCB concentrations of 100 mg/kg or less using the WES RSLS.
Results of the surface runoff tests conducted immediately after placement of
sediment in the lysimeter showed that contaminants were primarily in the
paniculate phase. Suspended solids concentrations were high (> 7,000 mg/f),
resulting in high unfiltered concentrations of contaminants. Copper was the
only contaminant exceeding the U.S. Environmental Protection Agency Acute
109
Chapter 5 Losses From Confined Disposal Facilities
-------
Water Quality Criteria for the Protection of Marine Aquatic Life in filtered
samples. Filtered PCB concentrations were statistically less than the criteria.
After 6 months of drying and aging, a hard crust formed that reduced the
erosiveness of the sediment. Results of surface runoff tests conducted
6 months after drying and aging showed that filtered cadmium, copper, and
zinc concentrations were not significantly different from unfiltered concentra-
tions, indicating that these metals were primarily present in soluble forms.
Filtered copper and zinc were statistically greater than or equal to the
U.S. Environmental Protection Agency Acute Water Quality Criteria for the
Protection of Marine Aquatic Life. Both unfiltered and filtered PCB concen-
trations decreased in surface runoff after drying and aging.
Simplified laboratory tests
The WES RSLS described previously requires substantial quantities of
sediment for testing and to properly simulate the physicochemical changes that
are associated with drying and oxidation, 6 months to complete a test. The
Indiana Harbor studies (Environmental Laboratory 1987) included investiga-
tion of laboratory batch extractions for predicting runoff quality from wet,
anaerobic dredged material and dry, oxidized dredged material. The tests for
wet, anaerobic dredged material involved serial dilution of suspended solids.
The tests for dry, oxidized dredged material included various short-term dry-
ing and chemical extraction procedures. The results for predicting wet, anaer-
obic dredged material runoff quality by solids dilution and predicting dry,
oxidized dredged material runoff quality by peroxide oxidation were promis-
ing. Additional testing and verification on a number of different sediments
were recommended.
110
Chapter 5 Losses From Confined Disposal Facilities
-------
6 Contaminant Losses for
In Situ Capping and Capped
Disposal
Background
General
In situ capping (ISC) is the placement of a covering or cap of clean mate-
rial over an existing deposit of contaminated sediment. Capping is also a
disposal alternative that can be considered when contaminated sediments are
removed as a cleanup measure. For the case of removal, capping is the con-
trolled accurate placement of contaminated material at an open-water disposal
site, followed by a covering or cap of clean material. For purposes of this
report, the term "contaminated" refers to material that is unacceptable for
unrestricted open-water disposal and the term "clean" refers to material that is
acceptable for such open-water disposal. Level bottom capping (LBC) is the
placement of a contaminated material on the bottom in a mounded configura-
tion and the subsequent covering of the mound with clean sediment. Con-
tained aquatic disposal (CAD) is similar to LBC but with the additional
provision of some form of lateral confinement (e.g., placement in bottom
depressions or behind subaqueous berms) to minimize spread of the materials
on the bottom.
Capping is considered an appropriate contaminant control measure for
benthic effects in the Corps dredging regulations (33 CFR 335-338) and sup-
porting technical guidelines (Francingues et al. 1985). An illustration of ISC,
LBC, and CAD is shown in Figure 31.
Capping, a technology for isolating contaminated material, was developed
as a control measure for contaminant effects on benthic organisms. The clean
material in a cap isolates benthic organisms that recolonize a site from the
contaminants in the material beneath the cap. The release of contaminants
into the water column is not generally viewed as a significant problem for
dredged material from most navigation projects. However, when capping is
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
CONTAMINATED DREDGED MATERIAL
(CLEAN SAND, SILT. ETC.)
CONTAMINATED DREDGED MATERIAL
Figure 31. Capping alternatives
112
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
considered as an alternative for sediment remediation, contaminant release to
the water column must be considered.
Design requirements for capping
Capping should not be viewed merely as a form of restricted open-water
disposal. A capping operation is an engineered project with carefully consid-
ered design, construction, and monitoring. The basic criterion for a success-
ful capping operation is simply that the cap thickness required to isolate the
contaminated material from the environment be successfully placed and
maintained.
Guidelines are available on planning and design concepts (Truitt 1987a, b),
design requirements (Palermo 1991a), site selection considerations (Palermo
1991b), equipment and placement techniques (Palermo 1991c),and monitoring
(Palermo, Fredette, and Randall 1992) for capping projects. These guidance
documents were developed primarily for capping projects associated with
navigation dredging; however, they are also applicable to capping associated
with sediment remediation to include ISC, LBC, and CAD projects. A cap-
ping guidance document is being prepared specifically for in situ subaqueous
capping of contaminated sediments that should be consulted when it becomes
available (Palermo et al., in preparation).
Influence of Capping Materials, Site, and
Operations
The nature of the material to be capped, the nature of the capping site, and
the dredging and placement equipment and techniques used will have direct
influence on the potential contaminant releases associated with capping. These
essential components of the design must be examined as a whole with compat-
ibility in mind.
A major consideration in compatibility is an acceptable match of equipment
and placement techniques for contaminated and capping material. For exam-
ple, if the contaminated material were mechanically dredged and released from
barges, the capping material could be similarly placed or could be placed
hydraulically. However, if the contaminated material were hydraulically
placed, then only hydraulic placement of the capping material may be appro-
priate due to the potentially low shear strength of the hydraulically placed
material.
Compatible scheduling of the contaminated material placement and capping
operation is essential. The exposure of the contaminated material to the envi-
ronment and need to allow consolidation of the contaminated material to occur
prior to cap placement must be balanced in scheduling both placement opera-
tions. Availability of equipment and funding and the possibility of equipment
113
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
breakdowns or other delays should be considered in determining if the capping
schedule is compatible with the contaminated material placement schedule.
Mechanisms for Contaminant Loss During Capping
For capping projects, the mass release is that total contaminant mass that is
not initially capped or that does not remain isolated by the cap. This defini-
tion implies that both short-term losses during contaminated and capping
material placement and long-term losses following completion of the construc-
tion of the cap must be considered.
Mechanisms for contaminant loss associated with capping therefore include
the following:
a. Water column during placement of contaminated material.
b. Resuspension during placement of cap.
c. Pore water expulsion during cap consolidation.
d. Long-term diffusion and advection.
e. Long-term bioturbation.
/. Long-term erosion.
For LBC and CAD, contaminated material is dredged, transported, and
placed at a capping site; therefore, losses for these components must be con-
sidered. It is anticipated that the majority of capping projects for sediment
remediation will be in situ. For ISC, there is no dredging or placement of
contaminated material and, therefore, no contaminant loss associated with
contaminated material placement. Resuspension of the contaminated material
and associated loss and long-term losses associated with diffusion, advection,
bioturbation, and erosion processes must be considered for ISC, LBC, and
CAD alternatives.
Water Column Contaminant Loss During Placement
Mass release of contaminants
Prediction of water column losses in terms of mass release for capping
during placement of the contaminated material for LBC and CAD alternatives
can be made using similar approaches as normally used for prediction of water
column releases for open-water disposal operations (USEPA/USACE 1992).
The approach taken is to determine contaminant concentrations associated with
both dissolved and suspended paniculate phases by standard elutriate testing.
114
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
Modeling of the fluid and suspended solids plumes is then used to predict the
losses.
Standard elutriate testing
The prediction of dissolved and particle-associated releases of contaminants
relies on the standard elutriate test. This test was developed in the early
1970s as a regulatory tool, and its utility and accuracy have been extensively
field verified (Burks and Engler 1978; Brannon 1978; Jones and Lee 1978).
In normal practice, the test is used as a predictor of dissolved contaminant
releases resulting from open-water discharge of dredged material for purposes
of comparison with applicable water quality criteria or standards, and to
develop an appropriate medium for conducting water column bioassays
(USEP A/US ACE 1992). If total concentrations of contaminants are measured
in the test, the results can be used in conjunction with modeling to calculate
mass release of contaminants associated with the suspended solids (Palermo
et al. 1989).
The standard elutriate test consists of the following steps as illustrated in
Figure 32:
WATER FROM
DREDGING SITE
SEDIMENT
80% BY VOLUME 20% BY VOLUME
(SETTLE FOR \
1HR J
SHAKE VIGOROUSLY IN
FLASK FOR 30 MIN.
\
)
CENTRIFUGATION OR
0.45 mm FILTRATION
)
CHEMICAL ANALYSIS
DISSOLVED CONCENTRATION
Figure 32. Standard elutriate test procedure
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
115
-------
a. Mix dredging site sediment and water to a sediment-to- water ratio
of 1:4 on a volume basis at room temperature.
b. Stir the mixture vigorously for 30 min with a magnetic stirrer. At
10-min intervals the mixture is also stirred manually to ensure com-
plete mixing.
c. Allow the mixture to settle 1 hr.
d. Siphon off the supernatant and centrifuge or filter (0.45 /*m) to remove
particulates prior to chemical analysis for dissolved contaminant
concentrations.
e. If particle-associated concentrations are desired, split the supernatant
immediately after siphoning into subsamples for dissolved and total
concentrations of contaminants and concentration of total suspended
solids.
The dissolved concentrations from the test are the predicted dissolved
concentrations in the discharge. The contaminant concentrations associated
with suspended solids is the difference between total contaminant concentra-
tions in whole water samples and dissolved contaminant concentration in the
filtered water samples (Equation 46 below).
r _ Ctotal ~ Cw (46)
~ ~~-
where
Cps = suspended solids contaminant concentration, mg/kg
Ctotal = whole water contaminant concentration, mg/l
Cw = dissolved contaminant concentration, mg/f
Cp = suspended solids concentration of elutriate sample,
It should be noted that Cw and Cs in the above equation are not necessarily
equilibrium concentrations. They could be equilibrium concentrations, but
equilibrium is not a necessary condition in the standard elutriate test.
Open-water disposal modeling
Computer models are available for predicting water column dispersion and
mixing (USEPA/USACE 1992 and Johnson 1990). The models also predict
the amount of material that would be lost to the water column during place-
ment. The use and limitations of the models along with theoretical discussions
1 1 R
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
are presented in detail in Johnson (1990). If barge release or hopper dredge
release is used, these models also indicate the initial spread of a single barge
load. This information is needed for evaluating mounding characteristics for
the material volume to be placed.
The models are available as a part of the Automated Dredging and Dis-
posal Alternatives Management System (ADDAMS) (Schroeder and Palermo
1990) and can be run on a microcomputer. ADDAMS is an interactive
computer-based design and analysis system for dredged material management.
The general goal of the ADDAMS is to provide state-of-the-art computer-
based tools that increase the accuracy, reliability, and cost-effectiveness of
dredged material management activities in a timely manner.
Model descriptions. The models account for the physical processes deter-
mining the short-term fate of dredged material disposed at open-water sites.
The models provide estimates of water column concentrations of suspended
sediment and contaminant and the initial deposition of material on the bottom.
Two of the models were developed by Brandsma and Divoky (1976) under
the Corps Dredged Materials Research Program to handle both instantaneous
dumps and continuous discharges. A third model that utilized features of the
two earlier models was constructed later to handle a semicontinuous disposal
operation from a hopper dredge. These models are known as DIFID (Dis-
posal From an Instantaneous Dump), DIFCD (Disposal From a Continuous
Discharge), and DIFHD (Disposal From a Hopper Dredge). Collectively, the
models are known within ADDAMS as the Open-Water Disposal (DUMP)
Models.
For evaluation of initial mixing for ocean disposal, the models need only
be run for the contaminant requiring the greatest dilution to meet the respec-
tive water quality criteria. A data analysis routine is contained in the models
for calculating the required dilutions and determining which contaminant(s)
should be modeled.
In all three models, the behavior of the material is assumed to be separated
into three phases: convective descent, during which the dump cloud or dis-
charge jet falls under the influence of gravity and the initial momentum of the
discharge; dynamic collapse, occurring when the descending cloud or jet
either impacts the bottom or arrives at a level of neutral buoyancy where
descent is retarded and horizontal spreading dominates; and passive transport-
dispersion, commencing when the material transport and spreading are deter-
mined more by ambient currents and turbulence than by the dynamics of the
disposal operation.
These models simulate movement of disposed material as it falls through
the water column, spreads over the bottom, and finally is transported and
diffused as suspended sediment by ambient currents. DIFID is designed to
simulate the movement of material from an instantaneous dump that falls as a
hemispherical cloud. Thus, the total time required for the material to leave
117
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
the disposal vessel should not be greater than the time required for the mate-
rial to reach the bottom. DIFCD is designed to compute the movement of
material disposed in a continuous fashion at a constant discharge rate. Thus,
it can be applied to pipeline disposal operations in which the discharge jet is
below the water surface or discharge of material from a single bin of a hopper
dredge. If the initial direction of disposal is vertical, either the disposal
source must be moving or the ambient current must be strong enough to result
in a bending of the jet before the bottom is encountered. DIFHD has been
constructed to simulate the fate of material disposed from stationary hopper
dredges. Here, the normal mode of disposal is to open first one pair of
doors, then another, until the complete dump is made, which normally takes
on the order of a few minutes to complete. DIFHD should not be applied to
disposal operations that differ significantly from the stationary hopper dredge
operations described above.
DIFID, DIFCD, and DIFHD model disposed dredged material as a dense
liquid. This model assumption will be satisfied if the material is composed of
primarily fine-grained solids. Thus, the models should not be applied to the
disposal of sandy material. A major limitation of these models is the basic
assumption that once solid particles are deposited on the bottom, they remain
there. Therefore, the models should only be applied over time frames in
which erosion of the newly deposited material is insignificant.
The passive transport and diffusion phase in all three models is handled by
allowing material settling from the descent and collapse phases to be stored in
small Gaussian clouds. These clouds are then diffused and transported at the
end of each time step. Computations on the long-term grid are only made at
those times when output is desired.
Model input. Input data for the models are grouped into the following
general areas: (a) description of the disposal operation, (b) description of the
disposal site, (c) description of the dredged material, (d) model coefficients,
and (e) controls for input, execution, and output.
Ambient conditions include current velocity, density stratification, and
water depths over a computational grid. The dredged material is assumed to
consist of a number of solid fractions, a fluid component, and a conservative
contaminant. Each solid fraction must have a volume concentration, a specific
gravity, a settling velocity, a void ratio for bottom deposition, and information
on whether or not the fraction is cohesive. For initial mixing calculations,
information on initial concentration, background concentration, and water
quality criteria for the constituent to be modeled must be specified. The
description of the disposal operations for the DIFID model includes position
of the disposal barge on the grid, the barge velocity, and draft, and volume of
dredged material to be dumped. Similar descriptions for hopper dredge and
pipeline operations are required for the DIFCD and DIFHD models. Coeffi-
cients are required for the models to accurately specify entrainment, settling,
drag, dissipation, apparent mass, and density gradient differences. These
coefficients have default values that should be used unless other site-specific
118
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
information is available. Appendix C - Table Cl lists the necessary input
parameters with their corresponding units. More detailed descriptions and
guidance for selection of values for many of the parameters is provided
directly on-line in the system.
Model output. The output starts by echoing the input data and then
optionally presenting the time history of the descent and collapse phases. In
descent history for the DIFID model, the location of the cloud centroid, the
velocity of the cloud centroid, the radius of the hemispherical cloud, the den-
sity difference between the cloud and the ambient water, the conservative
constituent concentration, and the total volume and concentration of each solid
fraction are provided as functions of time since release of the material. Like-
wise, the location of the leading edge of the momentum jet, the center-line
velocity of the jet, the radius of the jet, the density difference between mate-
rial in the jet and the ambient water, the contaminant concentration, and the
flux and concentration of each solid fraction are provided as functions of time
at the end of the jet convection phase in DIFCD and DIFHD.
At the conclusion of the collapse phase in DIFID and DIFHD, time-
dependent information concerning the size of the collapsing cloud, its density,
and its centroid location and velocity as well as contaminant and solids con-
centrations can be requested. Similar information is provided by DIFCD at
the conclusion of the jet collapse phase. These models perform the numerical
integrations of the governing conservation equations in the descent and col-
lapse phases with a minimum of user input. Various control parameters that
give the user insight into the behavior of these computations are printed before
the output discussed above is provided.
At various times, as requested through input data, output concerning sus-
pended sediment concentrations and solids deposited on the bottom can be
obtained from the transport-diffusion computations. With Gaussian cloud
transport-diffusion, only concentrations at the water depths requested are
provided at each grid point. The volume of each sediment fraction that has
been deposited in each grid cell is also provided. At the conclusion of the
simulation, the thickness of the deposited material is given.
For evaluations of initial mixing for ocean disposal, results for water col-
umn concentrations can be computed in terms of milligrams per liter of dis-
solved constituent or in percent of initial dredged material suspended phase
concentration. The maximum concentration within the grid and the maximum
concentration at or outside the boundary of the disposal site are tabulated for
specified time intervals.
Calculation of mass release
Estimation of both concentrations and volumes are required to compute
mass release. Estimation of concentrations using standard elutriate results as
described above is fairly straightforward. However, the estimation of fluid
119
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
and solids fractions released based on the model results requires a definition of
what is considered a release.
The mass release of dissolved contaminants can be determined from the
dissolved contaminant concentrations as defined by the standard elutriate test
and the total volume of water entrained during dredging and released during
the discharge. A conservative approach is to assume that the total volume will
be released (Palermo et al. 1989). The volume of the fluid fraction is depen-
dent on the in situ density of the sediment dredged and the volume of water
entrained during dredging. This is a conservative approach, especially for
mechanically dredged material discharged from barges, because a large por-
tion of the fluid fraction will descend to the bottom as interstitial water with
the solids and will be capped.
The mass release associated with the particle fraction is more difficult to
calculate. Several factors must be considered and several approaches can be
taken. The model results include an estimation of the total fraction of material
remaining in suspension as a function of space and time. The "footprint" of
the deposit of contaminated material can also be determined from both model
results for a single discharge and the anticipated evolution of the mound size
for the total volume of material to be placed, including the capping material.
One approach is to assume that all material remaining in suspension after a
given time period is released. The appropriate time period used can be deter-
mined by the frequency of discharges from the barge or hopper, current con-
ditions, and the disposal site size and anticipated size of the overall capped
mound or deposit, considering the total volumes placed. Time periods on the
order of 30 min have been used for such estimates (Palermo et al. 1989).
Another approach is to examine the total volume of solids deposited within the
anticipated footprint of the deposit to be capped and assume that all solids not
settling within that footprint will be released. In either case, the results of the
model should be carefully considered in making the estimates. Past field data
have indicated that only a small fraction (a few percent) of the total mass of
material will not quickly settle to the bottom and therefore could not be ini-
tially capped (Truitt 1986).
Based on the above considerations, the following steps should be followed
in calculating the mass release during placement of contaminated material:
a. From standard elutriate test, determine dissolved and particle-
associated concentrations for the open-water discharge.
b. Determine the volume of the water fraction of the discharge based on
predredging sediment water content and anticipated water entrainment
during dredging.
c. Calculate the total mass release of the dissolved fraction as the product
of the dissolved concentration and the volume of water released (for
1 20
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
pipeline discharges, the mass release is the product of the concentra-
tion, flow rate, and time duration of the discharge).
d. Determine the total mass of suspended solids considered a release
based on model results.
e. Calculate the mass of contaminants associated with the suspended
solids as the product of the particle-associated concentration and the
mass of solids released.
/. Calculate the total mass release as the sum of the dissolved and
particle-associated releases.
Water column control measures
If the total mass release to the water column during placement is unaccept-
able, control measures could be considered to reduce the potential for water
column effects or other dredging equipment and placement techniques, or use
of another capping site could be considered. Control measures could include
use of a submerged discharge point, submerged diffuser, tremie pipe, hopper
dredge pumpdown, or similar equipment (Truitt 1987b).
Resuspension During Cap Placement
Resuspension of contaminated material already on the bottom by impact of
discharges of capping material is a potential contaminant release mechanism
for ISC, LBC, and CAD alternatives. However, the design of caps (Palermo
1991a) normally requires an excess thickness of capping material to account
for inaccuracies in the placement process. The placement technique for the
cap must be carefully chosen to minimize displacement and mixing of the
contaminated and capping material. In general, the choice of capping mate-
rials and placement techniques is intended to result in a cap with an initial
density less than or equal to the deposit of contaminated material.
Resuspension of contaminated material during cap placement will be
located near the bottom and highly localized. Resuspended material should
settle back to the bottom almost immediately. The overall size of the deposits
laid down during capping and the gradual manner in which capping material is
placed tend to result in capping of material displaced in the early stages of the
capping operation. However, loss of contaminated material during cap place-
ment has not been extensively monitored, and there are no techniques avail-
able for preproject estimation of potential resuspension.
121
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
Losses During Consolidation
Contaminant losses during consolidation after cap placement may be impor-
tant especially for BLC and CAD. Pore water expressed through the cap will
result in the release of contaminants to the overlying water unless the cap has
sufficient sorption capacity to retain the contaminants. The release of contam-
inants via the expression of pore water through consolidation can be modeled
as a short-term advective process using the methods of the next section. For
organic contaminants, retention during consolidation is more likely if the cap
material contains significant organic matter.
Long-Term Contaminant Release Through Cap
Determine required cap thickness and exposure time
The cap must be designed to chemically and biologically isolate the con-
taminated material from the aquatic environment. Determination of the mini-
mum required cap thickness is dependent on the physical and chemical
properties of the contaminated and capping sediments, the potential for biotur-
bation of the cap by aquatic organisms, and the potential for consolidation and
erosion of the cap material. Laboratory tests have been developed to deter-
mine the thickness of a capping sediment required to chemically isolate con-
taminated sediment from the overlying water column (Sturgis and Gunnison
1988). These tests can also be performed in the presence of bioturbating
organisms (Brannon et al. 1985). An evaluation of the potential for coloniza-
tion of the capped site by bioturbating organisms and the behavior of those
organisms with respect to intensity and depth of burrowing must be made.
The minimum required cap thickness is considered the thickness required for
chemical isolation plus that thickness of bioturbation associated with organ-
isms likely to colonize the site in significant numbers.
The integrity of the cap from the standpoint of physical changes in cap
thickness and long-term migration of contaminants through the cap should also
be considered. The potential for a physical reduction in cap thickness due to
the effects of consolidation and erosion can be evaluated once the overall size
and configuration of the capped mound is determined. The design cap thick-
ness can then be adjusted such that the minimum required cap thickness is
maintained.
Most of the consolidation of the contaminated material will occur within a
few weeks of placement. Cap placement could be delayed an appropriate time
period to allow the majority of consolidation to occur. Such a delay also
holds advantage from the standpoint of resistance of the contaminated deposit
to displacement during cap placement. However, a delay exposes the contam-
inated material to the environment. An appropriate delay between contami-
nated material placement and capping must balance environmental exposure
122
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
with the engineering requirements of stability and the scheduling constraints of
the dredging required for capping.
There is potential for long-term migration of contaminants through the cap
due to consolidation of the contaminated material and diffusion and advection.
The techniques for evaluation of consolidation (Poindexter-Rollings 1990) can
be used to estimate the cap thickness potentially affected by the movement of
contaminated pore water. Theoretical models for evaluation of long-term cap
releases is discussed in the following section.
Models for long-term capping releases
The goal of capping is containment for a sufficiently long period of time
that natural degradation processes have the opportunity to render the contami-
nant harmless or to reduce the contaminant flux to levels that are protective of
ecological and human health. Due to the uncertainty associated with the rate
and existence of natural degradation processes, this discussion will assume no
irreversible fate processes and focus on the estimation of the undergraded
contaminant losses through a cap.
Potential long-term contaminant loss mechanisms for capped sediment are
essentially identical to the original uncapped sediments. The pore water trans-
port processes of diffusion and advection, perhaps enhanced by the presence
of colloidal particles in the pore water, are present. Particulates that remain
suspended can also enhance the transport of contaminants, but a cap should act
as an effective filter or scavenger of noncolloidal particulates. In addition,
and especially important for strongly sorbed contaminants, particle movement
processes such as erosion and deposition as well as bioturbation occur.
In the capped system, the bioturbating organisms at the original sediment-
water interface are buried, but recolonization of the upper cap layer occurs.
Over much of the capped depth, pore water processes such as molecular diffu-
sion and advection dominate transport processes. Erosion of the cap can
eliminate resistance to mass transfer provided by the cap by allowing deeper
penetration of the bioturbation layer. In the long-term models discussed in
this section, the cap is assumed stable or replaced as necessary to maintain
sufficient depth to avoid bioturbation of the original sediments. The effects of
slow depositional and erosional processes on contaminant transport through
caps are considered, but the effects of storm events on cap stability are not
included in the models discussed in this section. The long-term stability of a
cap can be assessed via the methods presented by Dortch et al. (1990) and
Maynord (1993).
Despite the similarity of transport processes in the capped and uncapped
sediment, the cap serves to reduce the net contaminant transport over the
uncapped situation as a result of the following:
1 73
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
a. Destruction of bioturbating organisms at original sediment-water
interface.
b. Increase of diffusion path length or advective path length before con-
taminants are transported to the water column.
c. Elimination of erosion at original sediment-water interface, at least
until erosion of cap.
d. Introduction of thermodynamic limitations due to elimination of particle
transport processes in the contaminated zone.
e. Retardation of pore water processes through the cap due to the pres-
ence of unfilled sorption sites.
In the following sections, processes affecting long-term cap effectiveness will
be discussed, and a quantitative analysis of these processes will be presented.
Molecular diffusion. Molecular diffusion is the process of random molec-
ular motion leading ultimately to equalization of chemical potentials every-
where within the system. In free water, the diffusive flux is written as
proportional to the concentration gradient in the water
NA = -DA2L (47)
A A2
where
NA = flux of contaminant A in free water, g/m2 sec
DA2 = diffusivity of A in water, m2/sec
Cw = dissolved concentration of A, g/m3
z = distance through water, m
The diffusion coefficient is of the order of 10~5 cm2/sec (10"9 m2/sec) in water.
The minus sign is needed in Equation 47 because contaminants diffuse from
regions of higher concentrations to regions of lower concentrations by random
molecular motion. The random motion of molecules that leads to diffusion
generally occurs at significant rates only within the pore spaces of the sedi-
ment or the overlying cap. Therefore, diffusivity must be corrected for the
available pore space in the media (e = porosity) and the fact that the diffusion
paths are not straight (T = tortuosity = actual path length/straight-line path
link). In a saturated, unconsolidated granular sediment, the tortuosity is
approximately e'1/3 (Millington and Quirk 1961) suggesting
124
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
N - - D w = - D e4/3
NA " A3 ~ A2
where
A^ = flux of contaminant A into cap, g/m2 sec
DA3 = effective diffusivity of A in sediment, m2/sec
Cw = water concentration of A, g/m3
z — distance into sediment or cap, m
DA2 = diffusivity of A in water, m2/sec
e = sediment porosity, m3 voids/m3 total volume
Flux is positive when the movement is toward positive z, that is, into the cap
or sediment.
Advection. Advection is a process associated with the bulk movement of
the pore water in response to pressure or head gradients in the sediment.
Advective processes should be especially important near the banks of rivers,
shores of lakes, and in estuarine systems subject to significant tidal variations.
In many regions, there is insufficient information on the permeability and
hydraulic gradient to adequately assess the advective contaminant transport. If
such information is available, however, the advective flux is written as
follows:
NA-UCW (49)
where
NA = flux of contaminant A, g m"2 sec"1
U = Darcy water velocity, m/sec
Cw = dissolved concentration of A, g/m3
The Darcy velocity used to define the advective flux is averaged explicitly
over the entire cross-sectional area of the medium and implicitly over some
volume. This averaging fails to identify the variations in velocity that occur
both within a pore and between adjacent pores in the medium. The variation
in velocities on the microscale results in additional mixing of the contaminant
above what would result from molecular diffusion alone. By analogy with
molecular mixing, microscale dispersive mixing is parameterized as follows:
125
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
(50)
where
NA = flux of contaminant A, gl m"2 sec
EA3 = effective dispersion coefficient in medium, m2/sec
Cw = dissolved concentration of A, g/m3
z = distance through water, m
Although the effective dispersion coefficient can be estimated from medium
properties, better estimates are obtained from laboratory contaminant transport
or tracer experiments that simulate field conditions.
The dispersion coefficient is often taken as approximately proportional to
the Darcy velocity. The constant of proportionality, the dispersivity, is
related to the characteristic size of microscale heterogeneities. For a homo-
geneous, granular medium, the dispersion coefficient is expected to be approx-
imately half of the particle diameter, that is
Ea3 = r U = -L U (5D
where
Ea3 = dispersion coefficient in medium, m2/sec
T — dispersivity, m
U = Darcy velocity, m/sec
dp = particle diameter, m
Very low advective velocities can control contaminant transport when
compared with diffusion. The importance of advection relative to diffusion
can be quantified by the Peclet number (Pe), which is defined
Pe = U L / DA3 (52)
126
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
where
U = advective velocity
L = transport length scale
D = effective diffusion coefficient
Since capping relies on reducing (by design) contaminant transport to diffu-
sion, evaluation of the Peclet number is very important. For example, the
effective diffusion coefficient in the cap is typically of order 10 cm2/year.
For a chemical isolation layer of only 10 cm, advection at only 1 cm/year is
approximately equal in importance to diffusion for transport. Due to the
potential importance of advective processes, the prevailing groundwater
velocities must be ascertained before confidence can be placed in the ability of
a cap to contain contaminants.
Facilitated transport. Advection, dispersion, and diffusion are pore water
processes that may be enhanced by the presence of colloidal particles in the
pore water. Colloidal organic matter in the pore water may be especially
important. Due to natural degradation processes, there typically exists col-
loidal organic carbon, for example, large molecular weight humic and fulvic
acids, at concentrations of the order 10 to 100 mg/t . Hydrophobic organic
contaminants can effectively sorb to this dissolved organic carbon in the same
manner that they sorb to organic carbon on the sediment surface. Since the
dissolved organic carbon (DOC) is mobile, however, the presence of colloidal
organic matter essentially increases the capacity of pore water to carry con-
taminants. The DOC moves at the velocity of the pore water and with a
diffusivity of the same order of magnitude as the free water diffusivity of the
contaminant.
If the partition coefficient between pure water and the colloidal species is
Kc , then the advective and diffusive flux for a contaminant can be written
NA = U (1 + KcCc} Cw - DA2^ (l + D'KeCe) (53)
where
NA = flux of contaminant A is direction of bulk flow, g/m"2 sec
U = Darcy velocity, m/sec
Kc = colloid - water partition coefficient of A, m3/g
Cc = colloid concentration in water, g/m3
Cw = dissolved concentration of A, g/m3
127
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
DA2 = diffusivity of A in water, m2/sec
D' = ratio of colloidal species diffusivity to DA2, m2/sec
e = sediment porosity, m3 voids/m3 total volume
z = distance into sediment, m
For hydrophobic organic species, Kc should be of the same order as the parti-
tion coefficient between water and sediment organic carbon, Koc. In addition,
Cc should be approximately defined by the DOC for hydrophobic organic
contaminants if the particulate organic carbon is effectively scavenged by the
sediment. Finally, the diffusivity of the colloidal species in water is likely to
be approximately the same as the diffusivity of the contaminant species, that
is, D' is approximately equal to one since almost all organic species exhibit a
water diffusivity of the order of 10~9 m2/sec. With these assumptions, Equa-
tion 53 can be written as follows:
"A = U (1 - KocCdoc) Cw - DA2S'\l + KCCC) ^ (54)
where
NA = flux of contaminant A is direction of bulk flow, g/m2 sec
U = Darcy velocity, m/sec
Kc = colloid - water partition coefficient of A, m3/g
Cw = dissolved concentration of A, g/m3
DA2 = diffusivity of A in water, m2/sec
e = sediment porosity, m3 voids/m3 total volume
Cc = colloidal species concentration, g/m3
z = distance into sediment, m
Equation 54 is based on equilibrium partitioning concepts and is, therefore,
primarily applicable to organic contaminants. Guidance on applying a modifi-
cation of equilibrium partitioning to metals is available in Chapter 4 in the
section on a priori prediction. However, there is no guidance available for
colloidal species that might sorb metallic or elemental species.
Slow deposition and erosion. Deposition and erosion processes move
contaminants by exposing contaminated pore water and by movement of
128
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
contaminants sorbed to the depositing or eroding particles. For a particle
deposition velocity Ud, the flux of contaminants by this process can be written
"A = UdCs = Ud (« + PbKd + eCcKc) Cw (55)
where
AT, = flux of contaminant A in free water, g/m2 sec
Ud = net deposition velocity, m/sec
Cs = local sorbed concentration of A, typically concentration of A in cap
material at cap-water interface, g/m3
e = local porosity, m3 voids/m3 total volume
pb = local bulk density, g/m3
Kd = solids-water partition coefficient, m3/g
Kc = colloidal-water partition coefficient, mVg
Cc = colloidal species concentration, g/m3
Cw = local dissolved water concentration of A, g/m3
The first term in parenthesis in Equation 55 is that portion of the flux
associated with the pore water movement. The third term in parenthesis
represents that portion of the flux associated with the colloidal motion. The
second term in parenthesis represents the movement of contaminants sorbed to
the depositing or eroding particles.
Bioturbation. Bioturbation is an effective means of moving dissolved and
sorbed contaminants near the sediment-water interface. Bioturbation is proba-
bly the most significant mechanism for chemical transport from noneroding
bottom sediments. For lack of a better estimation method, bioturbation fluxes
are often modeled as an effective diffusion process. For example, it has been
estimated that bioturbation has resulted in an effective particle diffusion coeffi-
cient of about 10 cm2/year in New Bedford Harbor (Thibodeaux 1990). This
is approximately a factor of 10 smaller than the estimated molecular diffusiv-
ity. Since bioturbation is a particle movement process, however, the ratio of
bioturbation to molecular diffusion is the order of DJ£JDA-i, for a contami-
nant with a sediment-water partition coefficient of the order of 104 I/kg (for
example, PCBs in Indiana Harbor sediment (Environmental Laboratory
1987)), bioturbation in this case is approximately 103 times more rapid than
molecular diffusion. The bioturbation flux, assuming that it can be repre-
sented by a diffusion model, can be written as follows:
1 29
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
(56)
where
NA = flux of contaminant A out of sediment, g/m2 sec
Db = effective bioturbation diffusion coefficient, m2/sec
Q = sorbed concentration of A, g/m3
z = distance into sediment, m
€ = sediment porosity, m3 voids/m3 total volume
pb = bulk density, g/m3
Kd = sediment- water partition coefficient, m3/g
Kc = colloidal-water partition coefficient,m3/g
Cc = colloid concentration in water, g/m3
Cw = dissolved concentration of A, g/m3
Elimination of the organisms at the original sediment-water interface is a very
effective means of reducing the migration of contaminants from the sediment
into the overlying water as well as an effective means of isolating the contami-
nants from bottom-dwelling organisms. Recolonization of the new sediment-
water interface, however, reduces the effective cap thickness. Bioturbating
species are limited to the upper sediment, and many species are limited to
aerated sediments in the upper few centimeters. Some species, however,
burrow deeply into the sediment, and the occurrence of these organisms may
require a deeper cap or elimination of the capping alternative in particular
areas. Assessment of this problem requires a survey of the type and density
of the organisms in a particular contaminated sediment area prior to remedia-
tion planning.
Combined process model. The combination of all of the processes
discussed above into a dynamic mass balance on the capped sediment allows
estimation of the contaminant flux through the cap. The transient accumula-
tion of the contaminant includes accumulation in the pore water, on the colloi-
dal fraction in the pore water, and in the sorption sites in the cap. If it is
assumed as before that sediment- water partitioning is reversible, instantaneous
and linear, the conservation equation for contaminant transport in the cap can
be written as follows:
130
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
where / is a retardation factor defined by
(57)
Rf = 1 + + C^ (58)
and
e = sediment porosity, m3 voids/m3 total volume
Cw = dissolved concentration of A, g/m3
t = time, sec
Ud = net deposition velocity, m/sec
f/ = Darcy velocity, m/sec
£c = colloid-water partition coefficient, m3/g
Cc = colloid concentration in water, g/m3
z = distance into sediment, m
Db = effective biotubation diffusion coefficient, m2/sec
DA2 = diffusivity of A in water, m2/g
D' = ratio of colloidal species diffusivity to DA2, m2/sec
pb = bulk density, g/m3
Kd = sediment-water partition coefficient, m3/g
Dividing Equation 57 by Rf gives
131
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
dt
+
m
Rf
D F1/3
UA2 e
dz
(59)
*f Rf
where V is the interstitial velocity, or U/e, and all other terms are as defined
for Equation 57. The significance of the retardation factor, Rf, in Equation 59
is evident from its appearance in the denominator of several terms in the
equation. As indicated by Equation 58, Rf is always greater than or equal to
one. Thus, retardation reduces the significance of the terms that /fy-appears
in, and thereby retards the effective velocity or diffusion of a strongly sorbing
contaminant. The effective velocity is the bracketed term on the left-hand side
of the equation while the effective diffusion coefficient is the bracketed term
on the right-hand side of the equation. Equation 59 assumes that the partition
coefficients and colloid concentration are not spacially dependent. Solutions
of Equation 59 can be used to define concentration gradients in caps or,
through the previously defined flux equations, determine contaminant fluxes at
any time out of capped material.
Use of Equation 59 requires determination of the indicated parameters and
an appropriate means of using these parameters to define concentration or
fluxes as a function of time or position. Porosity and bulk density are sedi-
ment or field parameters that are often measured or are available. Molecular
diffusivity is a chemical-specific property that is tabulated or for which esti-
mation methods are available (Lyman, Rheel, and Rosenblatt 1990; Reid,
Prausnitz, and Sherwood 1977). Net deposition velocities and effective bio-
turbation diffusivities are site specific and difficult to measure since field data
are often limited to a small number of samples over short time periods. The
time evolution of vertical contaminant concentration profiles in sediments is
needed before accurate estimates of bioturbation diffusion coefficients can be
made. Generally, groundwater gradients and hydraulic conductivities in the
vicinity of a stream or lake are not known with sufficient resolution to accu-
rately predict groundwater flow velocities directly. In most large lake sys-
tems, however, significant convective velocities are likely to be confined to
the nearshore environment. Finally, chemical partitioning data are chemical
and sediment specific, and accurate determination of these terms require labo-
ratory tests such as batch or continuous leaching tests as discussed in Contami-
nant Losses During Pretreatment. In the absence of specific laboratory
characterization of the contaminant partitioning, estimation techniques can be
employed for hydrophobic organic chemicals as discussed in Appendix B.
As previously indicated, in the absence of direct measurements, Cc and Kc
are approximated by the dissolved organic carbon concentration and Koc,
respectively, for hydrophobic organic chemicals. A priori estimation of Koc is
discussed in Appendix B. Dissolved organic carbon concentration is difficult
132
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
to estimate without data from laboratory leach tests. D', the ratio of the
colloidal diffusivity to the effective contaminant diffusivity in the medium, can
be estimated from information on the size of the colloidal matter or assumed
to be approximately equal to 1. Thus, all of the parameters in Equation 59
can be estimated from sediment or cap chemical-physical properties deter-
mined in laboratory testing or from field data.
Use of the parameters as defined by either field, laboratory, or predictive
estimation techniques to the estimation of concentrations or fluxes with or
without a cap requires numerical or analytical solution of Equation 59. Ana-
lytical solutions will be preferred here recognizing that simple physical sys-
tems amenable to analytical solution are as sophisticated as can normally be
justified by the precision of the input parameters. Consistent with this goal,
analytical solutions will be described for the following:
a. Advective transport through a cap.
b. Steady-state diffusive flux through a capping layer.
c. Diffusive flux through capping layer at any time.
d. Time to diffusive breakthrough.
e. Time to diffusive steady-state flux.
In each case, a cap is assumed to be placed on a contaminated sediment as
shown in Figure 33. The result of the capping process is a layer of thickness
L of initially clean capping material that isolates the contaminants from the
bottom dwelling organisms and slows their release back into the water col-
umn. The sediments will be assumed to be sufficiently contaminated that the
contaminant concentrations in the material below the original sediment-water
interface remains essentially constant. This assumption provides an upper
bound to the actual contaminant release rate. The total depth, L, of cap is
assumed to be composed of two layers, Leap, a layer in which advection or
molecular diffusion dominate, and LBio, a layer in which bioturbation is the
dominant transport process. As will be indicated later, the rate of contaminant
transport in the bioturbated layer is likely to be much greater than that through
the remainder of the cap. Therefore, the effective thickness of the cap is
essentially equal to the total cap thickness minus the bioturbation layer.
Significant advection is an indication that capping may not be an appro-
priate containment mechanism. For compounds that can be sorbed by the
capping layer, a cap will provide containment for long periods of time, even
in the presence of advection. If advection is the dominant transport process,
the contaminant migrates through the cap at a rate given by U/Rf. A break-
through time, or the time until contaminants are observed in the water above
the cap, can thus be defined as
1 33
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
WATER
BIOTURBATION
LAYER
Lcap
CAP-C
MOLECULAR DIFFUSION
CONTROLLED LAYER
LBio
CONTAMINATED
: SEDIMENT
Figure 33. Definition sketch for in situ capping losses
R
'b,adv
u
(60)
For highly sorbing compounds such as PCBs or PAHs, advective transport
through the cap is still orders of magnitude smaller than the groundwater flow
velocities as long as the cap retains some sorption capability. A sand or
gravel cap, however, will be relatively permeable and will exhibit little or no
sorption, resulting in rapid breakthrough if advective transport should occur.
Caps composed of fine-grained material containing organic carbon will be
both more sorptive and less permeable. In addition, the extra resistance to
flow posed by the presence of the capping layer is likely to divert ground-
water flows to regions other than the capped sediment. Finally, permeability
control can always be achieved in particular situations by placement of a low
permeability layer such as a bentonite-impregnated fiber mat that will reduce
the expected advective flows to very low levels. Thus, it is expected that the
sites most suitable for capping will have adequately low groundwater
134
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
velocities or can be modified to reduce groundwater flows. In these cases,
molecular diffusion is expected to be the primary transport process in the cap,
and the subsequent discussion of contaminant losses will focus on that process.
In the molecular diffusion layer, the effective cap height is Leap and not L.
Within this layer of the cap, that is below the bioturbation layer, bioturbation
is negligible. In the region Leap, therefore, Equation 58 becomes
dC,
w
dt
,1/3
Rf
d2C.
w
(61)
The bracketed term on the right-hand side of Equation 61 is the effective
diffusion coefficient in the cap. This term accounts for facilitated transport
and sorption. As indicated by Equation 61, the contaminant transport rate
through the cap is reduced if sorption occurs in the cap, that is, Rf > 1. For
hydrophobic organic contaminants, this suggests that a high organic content
cap should be chosen.
In the region LBio, bioturbation is expected to be a much more rapid trans-
port process for sorbing contaminants than molecular diffusion so that molec-
ular diffusion can be neglected, and Equation 59 becomes
(62)
Solutions of Equations 61 and 62 can be used to describe concentrations and
fluxes of contaminants from the cap. Crank (1975) and Carslaw and Jaeger
(1959) present solutions to equations of the form of Equations 61 and 62
under a wide variety of boundary and initial conditions. In the sections that
follow, selected solutions will be presented that describe contaminant flux
through a cap initially clean of contaminants overlying a contaminated sedi-
ment layer of essentially constant concentration.
The maximum release rate will occur after contaminants have penetrated
through the entire cap. Since the amount of contaminant in the original sedi-
ment is assumed constant, steady-state solutions to Equations 61 and 62 exist
that represent this upper bound flux. Steady-state forms of Equation 61
(molecular diffusion layer) are given by
0 = DM e"3 (l + D'C (63-a)
and for£>' = 1,
135
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
(63-b)
where
DA2 = diffusivity of A in water, m2/sec
e = porosity, m3 voids/m3 total
D' = ratio of colloidal species diffusivity to DA2, m2/sec
Cc = colloid concentration in water, g/m3
Kc = colloid-water partition coefficient, m3/g
Cw = water concentration of A, g/m3
z = distance up into cap, m
Cpw = pore water concentration of A, including colloidal bound, g/m3
Under steady conditions, all sorption sites in the cap are filled and no transient
accumulation occurs. As a result, the retardation factor, which represents this
transient accumulation, does not appear in Equations 63-a and 63-b.
Steady-state forms of Equation 62 (bioturbation layer) are given by
(64-a)
~ V " - "// dz2
and
0 =
<* CP» (64-b)
From Equations 63-b and 64-b, steady-state flux through the cap is given by:
Nss = Kov (CpW - C *
(65)
where
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
cap
c Cc)
,1/3
and
Kb = 0.036
D
A2
eRfDb
!0.8
Cl/3
-i
and
Nss = steady-state flux, g/m2»s
Kov = overall mass transfer coefficient, m/year
Cpw° = pore water concentration in original sediment, g/m3
C* = background water concentration above cap, g/m3
DA2 = diffusivity of A in water, m2/sec
e = porosity, m3 voids/m3 total
Kc = colloid-water partition coefficient, m3/g
Cc = colloid concentration in water, g/m3
Kb = benthic mass transfer coefficient, cm/year
A = surface area of cap, yd2
v = kinematic viscosity of water, cm2/sec
v = current speed above cap, m/sec
Sc = Schmidt number, dimensionless = v/DM
hd = effective depth of cap, diffusive layer depth, m
hb = depth of bioturbation layer, m
Equation 66 can be simplified by defining a coefficient R such that
e Rf
R =
(66)
(67)
KCCC
(68)
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
137
-------
so that Equation 66 becomes
T
L
Bio
DA2 e1'3 R Db Kb
(69)
Techniques for predicting the pore water contaminant concentration, Cw,
below the original sediment-water interface were previously discussed in the
section on leachate quality. If a low solubility chemical is present as a pure
phase in the original uncapped sediment, Cw is limited by that solubility. As
indicated previously, one advantage of the cap is that direct exposure of chem-
icals in a pure phase is eliminated, and the pore water processes that control
are thermodynamically limited in their capacity for contaminant transport.
Equation 65 is written with pore water concentration (dissolved plus col-
loidally bound) as the input variable. The dissolved plus colloidally bound
concentration is operationally defined as the dissolved fraction in that it is the
concentration that is measured after filtering the water. Thus the normally
available dissolved concentration contains both dissolved and colloidally bound
contaminant, and Equation 65 is the appropriate equation to use. In addition,
partition coefficients between sediment and water usually are measured by
employing the operational definition of dissolved. That is, the water concen-
tration predicted by such a partition coefficient would be total pore water
concentration or the sum of the truly dissolved and the colloidal contaminant,
and again Equation 65 would be the appropriate equation to use with that
concentration.
An equivalent equation could be written with truly dissolved concentration
as the input variable and modified definition of the overall mass transfer coef-
ficient to include facilitated transport. If only the truly dissolved concentra-
tion is used, that is, if empirical relationships from the literature are used to
estimate distribution coefficients, the pore water concentration is given by
Cpw = £ (1 + KCCC) (70)
and the retardation coefficient is as previously defined. As discussed in
Appendix B, Kd is given by
and
1 38
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
Kd
Kc K Koc ~ T-
(71-b)
If K - Koc and Cc is approximated by DOC, then
(72)
If pore water concentrations are estimated from sequential batch leach tests as
previously described, then there is no need to adjust for facilitated transport.
Leachate concentrations provided by this test include colloidally bound con-
taminant. Distribution coefficients obtained from sequential batch leach tests
also include the influence of colloids. Retardation factors obtained from
sequential batch leach tests, therefore, should not be corrected to account for
facilitated transport. In this case, Equations 63-b or 64-a should be used, and
the retardation factor in these equations becomes
Rf = 1 +
(73)
The steady-state flux given by Equation 65 is an upper bound to the actual
release rate. If significant sorption occurs in the cap, the time required to
reach steady state can be very long. Solution of the transient flux equation,
Equation 61, in the molecular diffusion layer of the cap suggests that the ratio
of the release rate from the top of the cap at any time to the steady-state rate
is given by (Thoma et al. 1993)
(t)
= 1 + 2
(-1)" exp
hi
(74)
where
RA(t) = release rate of contaminant, at time t, g/sec
RA(t-*oo) = release rate of contaminant, at steady state, g/sec
D^ = effective diffusivity, bracketed term Equation 59, m2/sec
hd = effective depth of cap diffusive layer, m
From this solution, the time required to achieve a breakthrough flux that is
0.05 percent of the steady-state flux is given by
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
139
-------
T = 0.54 _ (75)
b
and the time required to achieve 95 percent of the steady-state diffusive flux
through the cap without bioturbation is given by
r, = 3.69 ' (76)
where TW is the time required for the instantaneous flux to approach the
maximum value, that is, the steady-state flux given by Equation 74. Since the
effective diffusivity for a sorbing compound may be of the order of
10"9 cm2/sec, this suggests that it could take thousands of years to achieve the
steady-state release rate defined by Equation 65. It should be realized that
even the steady-state release rate is still orders of magnitude lower than the
release rate from the uncapped contaminated sediment.
The model equations presented have received experimental validation in
small laboratory test cells in which the release rate of trichlorophenol was
monitored (Wang et al. 1991; Thoma et al. 1993). Field demonstrations of
capping have been conducted, and preliminary evaluations of capping effec-
tiveness have been published (O'Connor and O'Connor 1983; Brannon et al.
1985; Truitt 1986b; Brannon et al. 1986). The information presented in these
evaluations is insufficient to determine the field validity of the in situ model
equations, primarily due to the long time required for measurable contaminant
migration. In addition, the model equations discussed for in situ capping
provide estimates of minimum losses because they do not account for losses
during placement and cap consolidation and erosion.
Long-term capping model summary. The general theoretical framework
for modeling long-term capping effectiveness was presented. The general
model includes the following transport processes: molecular diffusion, advec-
tion, dispersion associated with advection, low-order deposition/erosion
(excludes storm events), bioturbation, and sorption by capping material.
Simple model equations that neglect deposition/erosion, bioturbation, advec-
tion, and dispersion were presented. These model equations indicate that
hydrophobic organic chemicals in sediments can be isolated from the overly-
ing water column as long as the cap is stable, cap thickness is sufficient to
eliminate bioturbation, and advective transport is less than diffusive transport.
140
Chapter 6 Contaminant Losses for In Situ Capping and Capped Disposal
-------
7 Contaminant Losses During
Effluent and Leachate
Treatment
Background
After contaminated sediment has been removed by dredging, effluent and
leachate discharges may be generated during pretreatment, treatment, and
disposal operations. Effluent is generated during pretreatment by dewatering
processes, during hydraulic disposal in CDFs, during mechanical placement in
nearshore and in-water CDFs, and as a process waste stream during dredged
material treatment. Leachate is generated at pretreatment and CDFs as a
result of consolidation of dredged materials and infiltration and percolation of
rainfall.
Both effluent and leachate may be collected for treatment and/or disposal,
or may be allowed to dissipate to the surrounding soil and waters. This
chapter addresses contaminant losses associated with various treatment alterna-
tives for effluent and leachate and will not address potential losses associated
with release of untreated effluent and leachate. Untreated effluent and leach-
ate losses can be estimated using the predictive techniques discussed in
Chapter 4.
Leachate and effluent from a single source will contain essentially the same
contaminants, with the primary differences being the respective volumes gen-
erated, concentrations of contaminants, oxidation-reduction potential, and pH.
Assuming the effects of the variable loading conditions can be effectively
managed, process efficiency data are needed in order to estimate contaminant
losses for treatment processes applied to effluent and leachate. Process effi-
ciency is a function of initial contaminant concentrations, waste stream charac-
teristics, process design, and unit operation and maintenance. At best, ranges
in process efficiency can be estimated a priori. Bench- and pilot-scale testing
is required to determine treatment effectiveness for specific processes and
waste streams. In most cases, complete destruction of contaminants is not
feasible, and some contaminant loss will occur in process waste streams.
141
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
The contaminants identified at the areas of concern under the ARCS pro-
gram include PCBs, heavy metals, and PAHs. Other organic priority pollut-
ants have been identified, but are generally present at concentrations of less
than 1 mg/kg. Removal of suspended solids, organic contaminants, nutrients,
ammonia, oxygen-demanding materials, oil and grease, and heavy metals can
also be of concern for dredged material leachate and effluent. Three treatment
technology types may be needed as follows: organic chemical removal or
treatment, suspended solids removal, and heavy metals removal.
Contaminant Loss Estimation
Estimation of contaminant losses during effluent and leachate treatment is
based on a materials balance of the process treatment train. A process flow-
chart should identify waste streams through which contaminants can escape
treatment or control. An example is shown in Figure 34. Process flowcharts
can be developed from site-specific bench- or pilot-scale treatability studies or
from treatability studies conducted on similar wastewaters. Sediment sampling
and appropriate laboratory tests as described in Chapter 4 are necessary to
determine effluent and leachate characteristics and contaminant concentrations.
Information on effluent and leachate characteristics, anticipated effluent and
leachate flows, and treatment process efficiencies, is needed before treatment
process trains and flowcharts can be fully developed.
Aqueous treatability data are available for many potentially applicable
treatment technologies that can be used for a priori estimation of contaminant
losses. These data, while suitable for planning level assessments, treatment
process screening, and contaminant loss estimation, are not always suitable for
site-specific design calculations. For this reason, bench- and/or pilot-scale
treatability studies are usually needed to fully evaluate candidate treatment
technologies. Treatability studies should be conducted such that the informa-
tion needed to estimate contaminant losses is obtained in addition to the infor-
mation needed for full-scale design.
Sources of information on treatment efficiency include Cullinane et al.
(1986), Berger (1987), Corbitt (1989), and Averett et al. (1990). Emerging
treatment technologies can be found in the USEPA site technology profiles
(USEPA 1993b). In addition, computerized databases are available from the
USEPA Risk Reduction Engineering Laboratory (RREL) (USEPA 1992), the
Vender Information System for Innovative Treatment Technologies (VISITT)
(USEPA 1993c), and SEDiment Treatment Technologies Database (SEDTEC)
(Wastewater Technology Centre 1993). The RREL database contains
1,166 chemical compounds and over 9,200 sets of treatability data. It is
available in diskette form for MS DOS personal computers and is menu driven
and easy to use. Table 10 illustrates some of the information available on
treatment processes available in the RREL database on aqueous waste steams.
The data listed in Table 10 represent composite results for a variety of wastes
142
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
DREDGED MATERIAL
*- ATMOSPHERE
DISCHARGE
Figure 34. Example effluent/leachate treatment process flowchart
(domestic wastewater, industrial wastewater, synthetic wastewater, etc.),
scales of treatment (bench, pilot, and full), and contaminant concentrations
(low, medium, and high). Designers and planners should consult the RREL
database or other sources for more detailed information on specific treatment
performance. In the remainder of this chapter, selected treatment technologies
identified by Averett et al. (1990) are briefly examined for process basics and
information on treatment efficiencies. These technologies are listed in
Table 11. The list of treatment technologies in Table 11 is not exhaustive,
and designers of effluent and leachate treatment systems could consider other
treatment technologies.
Organic Treatment Technologies
Carbon adsorption
Process description. Carbon adsorption is an effective treatment process
for soluble organic compounds, and its use typically follows biological
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
143
-------
Table 10
Selected Removal Efficiencies for Aqueous Waste Streams (From
RREL Treatability Database (USEPA 1992))
Chemical
Arsenic
Cadmium
Copper
Chromium
Lead
Aroclor 1254
Aroclor 1260
Acenaphtnene
Benzo(ghi)perylene
Fluoranthene
Treatment Process
CHPT
FIL
CAC
CHPT
FIL
CAC
CHPT
FIL
CAC
CHPT
FIL
CAC
CHPT
FIL
CAC
API
SED
CHOX(CL)
CHOX(OZ)
PACT
CHOX(CL)
CHOX(CL)
CHOX(OZ)
PACT
Percent Removed
30-90 +
17
34-92
38-99 +
25 -49
0-80
27-99
0-75
19-95
0-99 +
47 - 90
19-95
0-99 +
21 -66
39-99
18
52
48
91
90
73
8 -44
99 +
77
Note: Composite information not intended for design calculations.
CHOX(CL): chemical oxidation using chlorine.
CHOX(OZ): chemical oxidation using ozone.
CHPT: chemical precipitation.
FIL: filtration.
CAC: chemically assisted clarification.
API: American Petroleum Institute oil/water separator.
SED: sedimentation.
PACT: powered activated carbon.
144
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
Table 1 1
Process Options for Effluent/Leachate Component Technology
Types (From Averett et al. 1990)
Metal Removal
Flocculation/coagulation
Ion exchange
Permeable treatment
Bed/dikes
Precipitation
Coagulation flocculation
Constructed Wetlands
Organic Removal
Carbon adsorption
Oil separation
Floating skimmers
Gravity separation
Coalescing plate separator
Chemical oxidation of organics
Ozonation
Resin Adsorption
Ultraviolet IUV) hydrogen peroxide
UV/ozonation
Constructed Wetlands
Suspended-Solids Removal
Chemical clarification
Granular media filtration
Membrane microfiltration
Constructed Wetlands
treatment or granular media filtration. Oil and grease concentrations greater
than 10 mg/t in the influent necessitate pretreatment in order to protect the
hydraulic and adsorptive capacity of the carbon. Air stripping may be utilized
for this, but adds significantly to the overall cost of the treatment process. Oil
skimmers and coalescing plate skimmers, discussed later, can be used to
remove oil and grease, but may not be effective in reducing oil and grease
concentrations to 10 mg/t without the attendant use of de-emulsifying
processes. Another alternative is to sacrifice the top layers of the carbon bed.
Wastewaters with insoluble oil and grease concentrations as high as 50 mg/t
have been successfully treated in this manner. Suspended solids concentra-
tions also influence process efficiency and column life. For fluids with a
viscosity near that of water, downflow columns are suitable for influents
containing suspended solids at concentrations of up to 65 to 70 mg/£. Upflow
columns can handle more viscous fluids, but require suspended solids concen-
trations less than 50 mg/t (Cullinane et al. 1986). High concentrations of
calcium carbonate or calcium sulfate will coat granular activated carbon which
cannot then be regenerated. This can be dealt with by pH adjustment or by
the addition of a scale inhibitor (Berger 1987).
Carbon adsorption processes will reduce BOD, COD, and TOC in addition
to specific organic compounds. In general, carbon adsorption is not as effec-
tive for polar organic molecules as it is for nonpolar organic molecules. Non-
polar organics are hydrophobic and as a result have high adsorption potential
and low solubilities. Low solubility humic and fulvic acids sorb readily and
may exhaust the carbon. Carbon adsorption is reportedly very effective in the
removal of PCBs, with tests resulting in levels less than 1 p.g/1 (Carpenter
1986). Shuckrow, Pajak, and Osheka (1981) report percent reductions of
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
145
-------
92.5 to 99.9+ percent reduction for PCBs. For multicontaminant systems,
competitive adsorption can reduce the removal rates of some compounds by
50 to 60 percent (Shuckrow, Pajak, and Osheka 1981).
Process waste streams. Process waste streams from carbon adsorption
units vary according to unit design. The waste streams common to all carbon
adsorption units are spent carbon and process effluent. Other potential waste
streams are offgases and backwash waters. Since contaminants are removed
by sorption to the carbon, spent carbon is the primary waste stream. Regen-
eration of spent carbon is usually accomplished thermally and may involve a
gas phase contaminant release. Losses associated with the process effluent
should by design be acceptable, that is, the treatment unit should meet given
performance standards. Performance standards can be met by additional
treatment if necessary.
Oil separation
Process description. Oily compounds foul the surfaces of exchangers,
sorbents, and filters diminishing process effectiveness and shortening the
useful life of the equipment. Oil and grease must be removed prior to ion
exchange, carbon adsorption, and filtration. Oil separation can be achieved
with continuous or batch processes. Continuous processes such as floating
skimmers and coalescing plate separators rely on gravity separation and
require very low flow rates to be effective (Corbitt 1989; Averett et al. 1990).
The effectiveness of oil separation methods varies with the nature of the oil
in solution, flow rate, temperature, and pH. Gravity separation can poten-
tially be very effective in oil removal if a process train is developed that is
appropriate for the characteristics of the fluid. Where the free oil concentra-
tion exceeds 1,000 mg/£, a separator must precede coalescing units in order to
prevent fouling with excess oil. Oil skimmers can potentially remove 99 per-
cent of free oil at the water surface, provided oil loading rates do not exceed
the capacity of the skimmer. Process efficiency will ultimately be determined
by the distribution of soluble and emulsified oil and the effectiveness of floc-
culants and de-emulsifiers.
Process waste streams. Process waste streams for oil and grease removal
technologies include removed oil and grease, process effluent, and gasses and
vapors. The oil and grease that is removed may contain significant amounts
of contaminants such as PCBs. For this reason, the oil and grease stream is
usually subjected to further treatment, such as incineration.
Oxidation
Process description. Chemical oxidation is based on the reaction of
chemical oxidants with wastewater constituents to transform and degrade
contaminants. Oxidants include chlorine, ozone (discussed separately below),
146
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
permanganate, peroxide, fluorine, and hypochlorite. Chemical oxidation can
be used for treatment of dilute influents containing oxidizable organics. It is
not suitable for complex waste streams, due to the nonselectivity of many
oxidants. Highly concentrated waste streams require large inputs of oxidizing
agents for this reason. Chemical oxidation is also not suitable for highly
halogenated organics. Its use has been reported for aldehyde, mercaptans,
phenols, benzidine, unsaturated acids, cyanide, certain pesticides, and as a
pretreatment to biological treatment for refractory compounds. It has limited
application for slurries, tars, and sludges (Kiang and Metry 1982). Incom-
plete oxidation can occur, with the potential for the formation of toxic inter-
mediate oxidation products.
Process waste streams. Process streams from chemical oxidation units are
limited to the process effluent and, in some cases, vapors. Losses associated
with the process effluent should by design be acceptable.
Ozonation
Process description. Ozonation is an oxidation process applicable to
aqueous streams containing less than 1 percent oxidizable compounds. Many
organic compounds and a few inorganic compounds are amenable to treatment
with ozone. Ozonation is especially useful for those compounds that are
resistant to biological treatment. Ozone is nonselective, oxidizing natural
organics as well as contaminants (Averett et al. 1990). Ozonation is not
suitable with sludges and solids. As with other types of chemical oxidation,
toxic end products sometimes result. Ozone is an aggressive oxidant, acutely
toxic and corrosive, requiring special handling, equipment, and safety mea-
sures. An incidental benefit to ozonation is the increase of dissolved oxygen.
Process waste streams. Ozone reactors are usually sealed reactors with
only the inlet, outlet, and ozone piping present. As such, the only process
waste streams for ozonation units is the process effluent. Losses associated
with the process effluent should by design be acceptable, that is, the treatment
unit should meet given performance standards. Additional treatment is usually
not necessary.
UV/hydrogen peroxide and UV/ozone
Process description. Hydrogen peroxide and ozone in combination with
ultraviolet (UV) light are effective in oxidizing a wide variety of chemicals.
Process efficiency varies with the target chemical(s) and general quality of the
water to be treated. Process efficiency is poorest with wastewaters that are
highly colored or turbid.
Process waste streams. UV/hydrogen peroxide and UV/ozone oxidation
units are usually sealed reactors with only the inlet, outlet, and oxidant
addition piping present. As such, the only process waste streams from
147
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
UV/hydrogen peroxide and UV/ozone oxidation units is the process effluent.
Losses associated with the process effluent should by design be acceptable.
Additional treatment is usually not necessary.
Resin adsorption
Resin adsorption is applicable for the removal of color due to organic
material and to high levels of dissolved organics (Cullinane et al. 1986). The
mechanism of removal is primarily sorption, and organics are inhibitory to the
function of ion exchange resins targeting other contaminants such as metals.
Performance data for resin adsorption are limited, -and highly variable.
Published efficiencies for dilute solutions containing PCB congeners range
from approximately 20- to 100-percent removal. Shuckrow, Pajak, and
Osheka (1981) reported 99-percent removal of PCBs at 100 ng/t by Amberlite
XAD-2. Other sources indicated similar efficiencies for this resin with PAHs.
Data for other resins and solution concentrations were not readily available.
Because of performance variability between resins and under different operat-
ing conditions, treatability studies are the most reliable method of determining
potential efficiency for a particular waste stream.
Process waste streams. Process waste streams from resin adsorption units
are similar to those from carbon adsorption units. Major waste streams are
spent resin and process effluent. Other potential waste streams that are design
and operation dependent are offgases and backwash waters. Since contami-
nants are removed by sorption to the resin, spent resin is the primary waste
stream.
Constructed wetlands
Process description. Constructed pollution abatement wetlands can be
designed to retain and degrade many pollutants, including toxic organic chem-
icals. Natural wetlands also potentially retain and degrade pollutants; but in
the context of remediation, discharge of effluent or leachate to a natural wet-
land is not anticipated. Constructed pollution abatement wetlands have been
primarily used in tertiary treatment of municipal wastewaters and for pH
adjustment of acid mine drainage (Hammer 1989). The mechanisms of
organic contaminant removal include adsorption, biodegradation, accumulation
by microbes, and, to a lesser degree, plant uptake.
Process waste streams. Constructed wetlands are open systems with many
contaminant migration pathways. They are also extremely complicated sys-
tems with many internal mechanisms for contaminant retention and degra-
dation. Loss pathways include volatile emissions, leachate seepage,
biotranslocation, and discharged waters. There are virtually no a priori and
no laboratory-scale procedures for estimating contaminant losses from
constructed wetlands. Mesocosm studies (pilot-scale wetlands) can be
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
conducted to obtain treatment process data needed for design (Rogers and
Dunn 1992; Doyle, Myers, and Adrian 1993) and to estimate losses. Limited
information on key wetland features, such as vegetative cover, vegetation
type, area flooded, hydraulic retention time, etc., with organic chemical treat-
ment process efficiency is available (Reed 1990; Phillips et al. 1993). Little
information is available on the removal of PCBs, PAHs, and similar chemicals
in constructed wetlands. A database on wastewater treatment using con-
structed wetlands (North American Wetlands for Water Quality Treatment
Database) is available from USEPA (USEPA 1994c). The database includes
178 sites and 203 separate systems. Most of the treatment information in the
database is limited to BOD and nutrients.
Suspended Solids Removal Technologies
Chemical clarification
Process description. Chemical clarification is utilized to enhance gravity
separation of suspended solids by the addition of chemicals that cause aggre-
gation of particles in solution. Organic polyelectrolytes are of primary inter-
est as the flocculent for use under the ARCS program. Synthetic flocculants,
while more expensive than natural inorganic compounds, require smaller doses
to achieve the same treatment level.
Schroeder (1983) conducted studies to verify earlier results obtained in the
use of polyelectrolytes and to develop guidelines in the design and operation
of chemical clarification facilities for dredged material slurries and super-
natant. As a result of these studies, all inorganic flocculants and all nonionic
and anionic flocculants were eliminated in preliminary bench-scale tests, leav-
ing 14 polymers that were tested on 0.84-, 1.26-, 1.69-, and 2.11-g/f
suspensions (suspended-solids concentrations representative of selected CDF
effluents). The more highly cationic and higher molecular weight polymers
were most effective in bench-scale tests.
Design of a system to achieve these treatment levels will be highly site and
sediment specific. Schroeder (1983) developed laboratory testing procedures
to facilitate determination of appropriate mixing intensity and duration, settling
time and volume requirements, and polymer dosages. In general, polymer
dosages are directly proportional to the turbidity to be treated, and inversely
proportional to the amount of mixing. A properly designed and operated
system can achieve average effluent suspended-solids concentrations on the
order of 50 mg/f under continuous operation. Results may be somewhat
variable due to the dynamic nature of the system.
Process waste streams. Several waste streams are possible with chemical
clarification systems depending on design. These waste streams include the
process effluent, leachate, volatile losses, and solids removed from the sec-
ondary settling basin. If the secondary settling basin is designed for storage
of solids as well as clarification, then there will be no contaminant losses
149
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
associated with removal and treatment/disposal of settled solids. Volatile and
leachate losses can be estimated using the techniques described in Chapter 4.
Leachate losses can be controlled by lining the settling basins. The process
effluent will likely be the major pathway for contaminant loss with most
designs, even those that do not include a liner. This loss can be reliably
estimated using data from test procedures and design calculations described by
Schroeder (1983).
Granular media filtration
Process description. Granular media filtration is a polishing step for
water that has been pretreated by settling or chemical clarification. The water
may be passed through permeable filter dikes or weirs, filter cells, or package
filters. Filter cells and sand-filled weirs are vertical flow filters that can be
replaced or regenerated when exhausted. Permeable dikes provide horizontal
flow filtration and are nonrenewable once clogged. Package filters typically
contain disposable cartridges that can be replaced when the solids loading
capacity has been reached.
Granular media for suspended solids removal include fine gravel, sand,
anthracite, and coal. Sand-filled weirs can remove 60 to 98 percent of sus-
pended solids, reducing the concentration to 5 to 10 mg/l for initial concen-
trations up to 1 g/f. Efficiencies up to 90 percent have been achieved for
concrete filter cells with sand and carbon filter media (Averett et al. 1990).
Process waste streams. Waste streams from granular media filters include
the process effluent, backwash water, spent media, and volatile emissions.
Volatile emissions can be estimated using the techniques described in Chap-
ter 4. Design equations developed by Krizek, FitzPatrick, and Atmatzids
(1976) can be used for a priori estimation of treatment efficiency for sus-
pended solids and particulate-bound contaminants. For low-maintenance
designs not requiring backwashing or media replacement, process water and
volatile losses are the two loss pathways of concern. Systems in which the
media is not contained in a chamber or vessel, such as porous dikes, may also
have a leachate pathway. Systems that require periodic removal of spent
media will have losses associated with the ultimate disposition of the spent
media.
Membrane microfiltration
Process description. Membrane microfiltration can be effective for sus-
pended solids concentrations of 10 to 5,000 mg/l, with the incidental benefit
of particle-associated contaminant removal (Averett 1990).
Process waste streams. Membrane microfiltration units produce two
process waste streams, process effluent and the filter cake. The filter cake
will probably contain most of the contaminant mass introduced into the unit.
1 50
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
Ultimate disposition of this material (landfilling or further treatment), there-
fore, is key to evaluating contaminant losses associated with membrane micro-
filtration. Spent membranes may also have to be considered. The process
effluent may contain dissolved chemicals that can be removed by further treat-
ment if necessary.
Constructed wetlands
Process description. Constructed pollution abatement wetlands can be
very effective in removing sediment particles. Sediment removal effectiveness
depends on sediment load and constructed wetland design and operation. The
keys to effective removal are providing hydraulic retention times and quies-
cent conditions sufficient for settling. Establishment of emergent vegetation
also plays an important role. Although the study of sedimentation of wetlands
has been somewhat limited, sedimentation has been extensively studied in
river, reservoir, and wastewater engineering. The design equations used for
detention basins provide a suitable basis for estimating solids losses from
wetlands constructed to treat effluent and leachate resulting from dredged
material treatment.
Process waste streams. Suspended solids releases through water control
structures is the primary mechanism for solids losses in constructed wetlands.
Constructed wetlands properly designed to remove suspended solids routinely
remove up to 90 percent of the total input (Reed 1990).
Metals Removal Technologies
Precipitation
Process description. Heavy metals can be precipitated from water as
sulfides or hydroxides with the addition of lime or sodium sulfide. Floccu-
lants can also be used to enhance agglomeration of precipitate particles and
resulting sedimentation. Chemical precipitation is most effective following
sedimentation and prior to filtration. Sulfides tend to be less soluble and more
stable over a broad pH range than hydroxides. Theoretically, metals can be
removed to their minimum solubility concentrations by adjusting the pH
according to the behavior of a specific metal ion. Where more than one metal
is present, more than one adjustment may have to be made, and a composite
pH at which all or several of the metals present approach their minimum
solubility is commonly used. Adequate process control can be difficult to
achieve in precipitation units if influent flows and concentrations vary widely.
Process waste streams. Process waste streams from precipitation systems
are similar to those from flocculation/coagulation systems. These waste
streams include process effluent, volatile emissions, and solids removed from
clarifiers. If the clarifier is designed for storage of solids as well as clarifica-
tion, then there will be no contaminant losses associated with removal and
151
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
treatment/disposal of settled solids. For systems that involve solids removal
from clarifiers, there may be contaminant losses associated with ultimate
disposition of precipitated solids. Since precipitation systems are usually
fabricated with steel or concrete, leachate is not a contaminant loss pathway
for these systems. Volatile losses can be estimated using the techniques
described in Chapter 4. Process effluent losses are best estimated from labo-
ratory or pilot treatability studies.
Flocculation/coagulation
Process description. Of the two basic flocculants used to treat dredged
material effluent, inorganic compounds and cationic polyelectrolytes are the
most promising for freshwater slurries. Cationic, anionic, and nonionic poly-
electrolytes are all potentially effective for use with saltwater slurries (Averett
et al. 1990). As discussed previously in this chapter, suspended solids
removals of 84 to 95 percent were achieved in field trials using polyelectrolyte
flocculants. Given that heavy metals tend to associate with fine particles,
metals-removal efficiencies are likely to be similar to suspended solids
removal efficiencies. Flocculation added following precipitation treatment
would remove precipitates formed from the soluble metals fractions as well.
Process waste streams. Several waste streams are possible with
flocculation/coagulation systems, depending on design. These waste streams
include process effluent, leachate, and volatile losses. In addition, there may
be contaminant losses associated with ultimate disposition of solids removed
from clarifiers. If the clarifier is designed for storage of solids as well as
clarification, then there will be no contaminant losses associated with removal
and treatment/disposal of settled solids. Leachate losses will be negligible
from fabricated systems using steel or concrete. Earthen basins as clarifiers
will have a leachate pathway that can be minimized or eliminated using a
liner. Volatile and leachate losses can be estimated using the techniques
described in Chapter 4. Process effluent losses are best estimated from labo-
ratory or pilot treatability studies.
Ion exchange
Process description. Of the three major operating modes (fixed-bed con-
current, fixed-bed countercurrent, and continuous countercurrent), the fixed
bed countercurrent system is most common (Cullinane et al. 1986). Use of a
hydrogen exchange resin facilitates removal of anions, and the hydroxide form
facilitates removal of cations. For a mixed waste, resins in series targeting
first the organics (polar and nonpolar resins) and then the ionic species (cat-
ionic and anionic resins) are effective (Cullinane et al. 1986). Ion exchange is
valuable because of the selectivity exhibited by exchange resins (Corbitt
1989). This selectivity varies with ionic strength, the relative concentrations
of ions in solution, and to a lesser extent temperature and other factors.
Natural ion exchange mediums include clay, zeolites, sulfonated coal, and
1 52
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
peat. Synthetic resins have the advantage of controllable properties and high
capacity.
Process waste streams. Process waste streams from ion exchange resins
are spent resin and process effluent. Other potential waste streams that are
design and operation dependent are offgases and backwash waters. Since
contaminants are removed by ion exchange with a resin, spent resin is the
primary waste stream. Depending on the ultimate disposition of spent resin,
there may be losses associated with disposal of this material. Losses associ-
ated with the process effluent should by design be acceptable. These losses
can be controlled by additional treatment if necessary. Data from bench- or
pilot-scale treatability studies are needed for design and estimation of contami-
nant losses.
Permeable treatment beds/dikes
Process description. Permeable treatment beds and dikes were previously
discussed under suspended solids treatment technologies. Under optimum
conditions, filtration through these structures will remove 60 to 98 percent of
the suspended solids and sediment-bound contaminants (Cullinane 1986).
They may be constructed using limestone, crushed shell, activated carbon,
glauconitic green sands (zeolites), or synthetic ion-exchange resins at the core
to effect ion exchange or precipitation reactions in addition to simple filtra-
tion. Permeable treatment beds and dikes are capable of handling suspended
solids concentrations up to 1 g/t (Averett et al. 1990).
Process waste streams. Process waste streams for permeable treatment
beds and dikes are the same as previously discussed under suspended solids
treatment technologies.
Constructed wetlands
Process description. As previously discussed, constructed pollution abate-
ment wetlands are capable of removing a wide spectrum of waterborne pollut-
ants, including metals. Metals can be immobilized in constructed wetland
soils and sediments by biologically mediated reduction-oxidation (redox) and
pH reactions. Microbes in constructed wetlands soils and sediments utilize
available electron acceptors (oxygen, nitrate, ferric iron, sulfate, manganic
manganese, and carbon dioxide) to accomplish electron transfer reactions
required for obtaining energy from substrates (Turner and Patrick 1968). In
this process, pH is raised or lowered depending on starting conditions to near
neutral. Coupling of oxidation-reduction reactions with pH is a chemical
thermodynamic requirement for these reactions (Ponnamperuma 1972). Many
metals are relatively insoluble at near neutral pH and low redox potential.
Aerobic (high redox), acidic wastewaters introduced as subsurface flow to
constructed wetlands is neutralized with concomitant reduction in dissolved
metals. This basic principle has been effectively used to treat acid mine
1 53
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
drainage at numerous sites (Hammer 1989). To date, however, sufficient data
are not available for development of design equations or contaminant loss
estimation algorithms.
Process waste streams. Loss pathways include leachate seepage, biotrans-
location, and discharged waters. Discharged waters probably represent the
major loss pathway for metals. Wetlands constructed to process wastewaters
from mining activities vary widely in their metals removal efficiencies
(Phillips et al. 1993). Wetlands can be very effective in removing metals
(removal efficiencies greater than 90 percent) or can be completely ineffective.
Mesocosm studies (pilot-scale wetlands) can be conducted to obtain treatment
process data needed for design (Rogers and Dunn 1992; Doyle, Myers, and
Adrian 1993) and to estimate losses.
Summary
Treatability data needed for screening candidate treatment processes are in
some cases difficult to locate depending on the contaminants and treatment
processes of interest. Sources of information include Cullinane et al. (1986),
Averett et al. (1990), Corbitt (1989), Berger (1987), USEPA (1993b), and
Wastewater Technology Centre (1993). In addition, treatment technology
databases are available that provide information on treatment process perfor-
mance (USEPA 1992; USEPA 1993c; USEPA 1994c). Process treatment
efficiencies are usually given in terms of percent of contaminant removed. In
some cases, this is a function of initial concentrations of contaminants. From
percent removal data, planning level assessments of contaminant losses during
effluent and leachate treatment can be made.
Sediment sampling and appropriate bench-scale testing are necessary to
determine effluent and leachate characteristics and concentrations of contami-
nants present. From this information and information on expected flow,
candidate treatment processes for effluent and leachate can be evaluated in
bench-scale treatability studies. Treatment efficiencies and contaminant con-
centrations in process streams can be calculated on a case-by-case basis once
site-specific treatability data are available.
Treatability studies are considered to be a requisite part of any treatment
design activity. The chemical and physical interactions of waste components
and treatment processes require careful evaluation for effective implementation
of any treatment program. Attention to design and scale-up principles includ-
ing consideration of process control is a key element in achieving optimum
removal efficiencies and minimum contaminant releases.
1 54
Chapter 7 Contaminant Losses During Effluent and Leachate Treatment
-------
8 Contaminant Losses for the
No-Action Alternative
Background
The no-action alternative, as used in this report, describes an assessment of
current contaminant concentrations in sediments at a site and of potential
danger that may occur in the future if no remedial action is taken. The
assessment assumes that the natural events expected in a water body will be
allowed to run their course with no changes made in the water body manage-
ment plan (no changes in loads, dredging practices, etc.). The no-action
alternative also may be referred to as a baseline exposure assessment because
it serves as a basis from which to compare all action alternatives. With this
baseline, the relative benefits of remediation programs can be compared, and
the time required for the system to cleanse itself can be estimated.
Procedures for developing a no-action alternative
The general steps in establishing a no-action alternative are described
below. These steps are modified from guidance provided for conducting
remedial investigations and feasibility studies under the Comprehensive Envi-
ronmental Response, Compensation, and Liability Act (CERCLA) (USEPA
1989).
Step 1. The first step in the assessment of the no-action alternative, or
baseline exposure assessment, is identification of potential pathways by
which contaminants may migrate from a source to a point of contact that is
considered hazardous to humans or to terrestrial or aquatic life. This
assessment includes identifying the mechanisms affecting the release of the
chemical (e.g., from contaminated sediments) as well as the processes that
may affect the environmental transport of the chemical (e.g., via sediment
resuspension and food chain uptake). This identification step serves to
focus the assessment on critical exposure pathways.
Step 2. Once the source(s) and release mechanisms have been identified
for contaminants in surface waters, the no-action exposure assessment then
155
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
turns to an analysis of the environmental transport and transformation of
the contaminants of concern. This fate analysis considers the potential
environmental transport (e.g., surface water and groundwater); transforma-
tion (e.g., biodegradation, hydrolysis, and photolysis); and transfer mecha-
nisms (e.g., sorption and volatilization) to provide information on the
potential changes in the magnitude and extent of environmental
contamination.
Step 3. Next, the actual and potential exposure points for receptors (e.g.,
humans and aquatic life) are identified. As part of this evaluation, a
reasonable maximum exposure scenario should be developed that reflects
the type(s) and extent of the exposure that could occur based on the likely
or expected use of the site (or surrounding areas) in the future.
Step 4. Information developed in the first three steps then is integrated to
produce quantitative and/or qualitative estimates of the expected exposure
level(s) from the actual or potential release of contaminants from the site.
Step 5. A final step in the assessment is to establish the uncertainty associ-
ated with projections of contaminant fate. This uncertainty assessment may
be both qualitative and quantitative.
Mathematical models are commonly used as the integrating tools to provide
estimates of the expected exposure level(s) under future conditions. Configur-
ing contaminant fate and transport models to provide predictions requires the
projection of environmental and physical/chemical conditions into the future.
Because of the persistence of some chemicals, these modeling projections may
extend to 30 years or longer.
The uncertainty associated with future projections severely complicates the
identification of the reasonable maximum exposure scenario and the use of
contaminant transport and fate models in the evaluation of the no-action alter-
native. Application of mathematical models over the long time periods
required for the assessment of persistent chemicals is particularly difficult.
Because the response time of some chemicals is on the order of 20 to
100 years, the no-action alternative modeling scenario would have to be simu-
lated for that period of time. This introduces a level of uncertainty on how
projections are made. For example, an assessment of flows could be made
using the period of record flows for the simulations. However, the historical
flow pattern may not be a good estimator of flow conditions for the future.
The system could experience a major flood, the equivalent of which was not
represented in the period of record. Alternatively, a truly stochastic approach
could be used based on the historical distribution of hydrology. However, a
completely stochastic approach is usually not feasible unless relatively simple
models of contaminant transport and fate are used, due to the computational
burden imposed by complex models. Different approaches used in the model-
ing scenario to evaluate the no-action alternative may produce different esti-
mates of the time-to-recovery or potential exposure levels in the future.
156
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
Because most persistent contaminants in aquatic systems are associated with
sediments, they are moved or dispersed in association with major sediment
resuspension events. Properly accounting for these events is often very
important. For example, a 100-year flood event could be responsible for
movement of 90 percent of the total contaminants, all in the course of a few
days, with the contaminated sediments either being exposed or buried. Trying
to account for the effects of large events is difficult. Seldom is there enough
information to allow for a complete analysis of these infrequent conditions.
Not only must the flow and sediment transport be predicted for the future,
but future land uses and exposure pathways must be considered during the
evaluation of the no-action alternative. The determination and the ultimate
acceptance of the no-action alternative would be based on the reduction of
contaminants in the water body and the subsequent reduction in the associated
risk of exposure. This no-action reduction can be used as a comparison of the
effectiveness of some proposed remedial action plans, where the calculated
risks can be compared with that of the baseline risk assessment.
Levels of study complexity and uncertainty
The level of effort required in the analysis of the environmental fate and
transport of contaminants in the no-action assessment depends largely on the
complexity of the site. The goal is to gather sufficient information to ade-
quately and accurately characterize the potential exposure from the site, while
at the same time conducting the study as efficiently as possible. Factors that
may affect the level of effort required include (USEPA 1988): (a) the num-
ber, concentration, and types of chemicals present and the areal extent of the
contamination, (b) the quantity and quality of available supporting data, (c) the
number and complexity of the exposure pathways (including the complexity of
release sources and transport media), and (d) the required precision of analy-
ses, which in turn depends on site conditions.
Evaluation of the no-action alternative usually requires the use of
hydrodynamic/sediment transport and contaminant transport and fate models.
However, the level of complexity of the modeling study may vary for the
reasons cited above. There are basically three levels in which a no-action
alternative can be conducted:
a. Screening Level—A. simplified modeling method or analytical equations
can be used to give rough estimates of contaminant mobility and con-
centrations under a set of conditions. This level is useful in addressing
broad management questions over long time periods.
b. Descriptive Modeling—A contaminant transport and fate model could
be set up on the water body using flows derived from historical records
and sediment transport derived from sedimentation records. This
approach bypasses the use of the hydrodynamic and sediment transport
model, but still provides insight in how the contaminant will react over
1 57
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
long periods of time under variable flow and sediment transport
conditions.
c. Fully Predictive Modeling—A hydrodynamic and sediment transport
model could be utilized in predicting the flows and sediment transport
for the period of record. This information would then be linked with
the fate and transport model.
Most large water bodies in the United States have U.S. Geological Survey
gauging stations or National Oceanic and Atmospheric Administration water
elevation gauges from which one can obtain measured continuous flow or
water surface elevations for the period of record required for the modeling
study. This type of information can be used in all three levels of modeling.
These data can be analyzed to determine the variability of the water move-
ment, and a probability distribution function (PDF) can be generated. PDFs
are used to determine the probability of different flow regimes occurring that
might cause major scour events. These major events and their associated
probabilities can be incorporated into the no-action alternative modeling study.
Degradation processes are known, or can be estimated reasonably well, for
most contaminants. Values for these processes would be entered into the
contaminant transport and fate model and used throughout the simulation
period. However, the site-specific parameters used to describe the degrada-
tion processes are usually determined using data available only over relatively
short time periods in comparison to the time over which the no-action alterna-
tive will be evaluated.
The determination of the exposure pathway of concern will dictate the
spatial and temporal resolution needed in the contaminant transport and fate
model. If the only interest is in reduction of downstream loadings, large
spatial compartments and temporal information may be adequate.
The processing and averaging of data may affect the conclusions that
result from the evaluation of the no-action alternative. To illustrate, down-
stream contaminant concentrations using mean monthly flow data versus daily
flow data are compared in Figure 35. As illustrated, the differences between
the calculated downstream concentrations are minimal. However, if the expo-
sure pathway is bioaccumulation of the contaminant in fish, the spatial and
temporal prediction to model may be very important. If this information was
taken a step further and used in a bioaccumulation study, the arbitrary selec-
tion of mean monthly flow data over daily flow data could lead to differences
in the predicted contaminant concentration in the biota. Figure 36 shows the
predicted contaminant concentrations in fish for a heavy organic-like PCB.
Although there is a difference in the predicted fish concentration, the error is
not large compared with the predicted exceedance of the U.S. Food and Drug
Administration (FDA) action limit. But, in the case of a light organic illus-
trated by Figure 37, the selection of the flow criteria can have an impact on
the remediation decisions for the water body. In this case, the error is signifi-
cant compared with the predicted exceedance of the FDA action limit.
•1 C Q
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
0.16
0.14
I
O 0.12
F
o
I
t-
1
s
0.08
0.06
0.04
0.02
LEGEND
DAILY FLOW
MEAN MONTHLY FLOW
50
100
150
200
TIME, DAYS
250
300
350
400
Figure 35. Daily versus mean monthly contaminant concentrations
14
12
I
Q.
z- 10
o
I <
tn oc
C £ 8
O HI
UJ O
~
111 >— 6
OC z
0- <
P 4
O
O
-It
50
BASED ON MEAN MONTHLY
FLOW AND CONCENTRATION
BASED ON DAILY FLOW
AND CONCENTRATION
100
150
200
TIME, DAYS
250
300
350
400
Figure 36. Heavy organic bioaccumulation: mean monthly versus daily flow and contami-
nant concentrations
Chapter 8 Contaminant Losses for the No-Action Alternative
159
-------
o
*<
0.26
0.24 -
0.22 -
O 0.20
"- z
O UJ
111 U
O O
IU I-
o: z
a. <
z
5
O
o
0.18
0.16 -
0.14
0.12
0.10
0.08
BASED ON MEAN MONTHLY
FLOW AND CONCENTRATION
BASED ON DAILY FLOW AND
CONCENTRATION
100
150 200 250
TIME, DAYS
300
350
400
Figure 37. Light organic bioaccumulation: mean monthly versus daily flow and
concentration
Modeling the No-Action Alternative
The application of mathematical models to conduct a no-action alternative
is a multistep process. The steps for conducting a modeling study are
illustrated in Figure 38. Water and sediment transport are first predicted so
that this simulated information can be used by the contaminant transport and
fate model. Next, the hydrodynamic and sediment transport predictions are
used along with the estimates of contaminant loadings due to nonpoint/point
source loadings to predict changes in chemical concentrations in water and
sediments. This gives time-variable contaminant concentration profiles for
sediments and water column that can be utilized by bioaccumulation/food
chain models to predict contaminant body burden for fish.
Hydrodynamic models
To effectively predict the dissolved concentration of a contaminant, it is
important to characterize the transport of water within the system. The vari-
ability and distribution of water column contaminant concentrations can often
be largely explained by water transport alone. Water transport models are
160
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
MODELING FRAMEWORK
HYDRODYNAMIC
MODEL
SEDIMENT
TRANSPORT
CONTAMINANT
TRANSPORT
LOADING
STUDY
FOOD CHAIN
MODEL
Figure 38. Steps in modeling no-action alternatives
based on a balance of water mass and, for hydrodynamic models, a balance of
water momentum, which, like mass, is also a conservative property.
Characterization of water transport may be descriptive or predictive,
depending on the modeling approach. In a descriptive approach, flow patterns
are measured directly or inferred from measurements. The descriptive
approach is often adequate where the system is very simple (hydraulically) or
where only long-term, relatively crude estimates of water transport are
required.
Hydrodynamic models are used to predict changes in volumes, depths, and
velocities in response to changes in upstream flows, downstream flows, water
surface elevations, or bottom morphometry. Hydrodynamic models can be
used to predict flows for periods where direct measurements are not available.
Hydrodynamic models also may be used to estimate changes in flows that may
occur under future conditions, such as in evaluating the effects of changes in
dredging patterns.
Sediment transport models
Adequately characterizing the movement of sediments is a critical step in
the assessment of the no-action alternative. There are two primary goals of
the sediment transport component: (a) to predict the movement of the sedi-
ments themselves in order to estimate changes that may occur in patterns of
erosion, deposition, and transport, and (b) to estimate the transport of the
paniculate contaminant mass. Sediment transport models are based on a
balance of sediment mass. As with water transport, sediment transport may
be described or predicted in mass balance studies. The descriptive approach
has proven useful in providing crude estimates of the effects of sediment
transport on contaminant distributions. However, sediment transport is a very
Chapter 8 Contaminant Losses for the No-Action Alternative
161
-------
dynamic process, and the assumption of steady-state solids behavior is a gross
simplification.
In predictive sediment transport models, resuspension and transport are
computed using the output of a hydrodynamic model and the characteristics of
the sediments. The type of sediments of importance in contaminant studies
are cohesive sediments (e.g., silts and clays) rather than noncohesive sedi-
ments (e.g., sands). The sediment transport model is used to predict changes
in suspended solids concentrations, changes in sediment resuspension and
deposition, and changes in the structure of the sediment bed. As with hydro-
dynamic models, sediment transport models can be used to interpolate among
existing measurements or to estimate sediment transport for conditions for
which data are not available. The majority of sediment transport occurs under
extreme (rare) events, such as storms on lakes and large run-off events in
rivers. Since data are often not available for these rare events, sediment
transport models can be used to estimate transport under these conditions.
For example, they may be used to estimate whether contaminated sediments
may be buried or exposed under these conditions. This information can be
used in the evaluation of remedial actions as well as the no-action alternative.
Sediment transport models may also be used to evaluate the impact of remov-
ing or immobilizing sediments on subsequent erosion and deposition patterns.
For example, if sediments are removed from a particular area, sediment trans-
port models may be used to estimate how long it may take for the area to fill
in as well as to predict changes that may occur in deposition and erosion
areas. Table 12 suggests several hydrodynamic and sediment transport models
that could be implemented in a no-action alternative modeling study. This
table is restricted to models that are in the public domain. In addition to the
models listed below, a wide variety of models are available in the private
sector that may be suitable for use in the evaluation of the no-action
alternative.
Table 12
Suggested Hydrodynamic and Sediment Transport Models
Hydrodynamic and Sediment Transport Models
Name
SED-3D
SED-2D
HEC-6
RIVMOD
DYNHYD
RMA
CE-QUAL-RIV1
DAM BREAK
Source
USEPA, Athens
USEPA, Athens
HEC, U.S. Army
USEPA, Athens
USEPA, Athens
WES, U.S. Army
WES, U.S. Army
US NWS
Dimension
3-D
2-D
1-D
1-D
1-D
2-D
1-D
1-D
Sediment
Transport
Y
Y
Y
Y
N
Y
N
N
Cohesive
Sediments
Y
Y
Y
N
N
Y
N
N
Linked w/WQ
Models
N
N
N
Y
Y
N
N
N
162
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
SED-3D.
General description: SED-3D is a circulation, sediment dispersion, resus-
pension, and deposition model for far-field transport in lakes, estuaries,
coastal areas, and other water bodies. It employs approximate second-order
closure scheme (Sheng and Eliason 1991).
Capabilities and strength: SED-3D can be used to simulate water flow
and sediment transport in various water bodies under the forcing of winds,
tides, freshwater inflows, and density gradients, as influenced by Coriolis
acceleration, complex bathymetry, and shoreline geometry. The model can be
run in a fully three-dimensional mode, a two-dimensional vertically integrated
x-y mode, or a two-dimensional x-z mode. The model contains a free sur-
face. A simplified second-order closure model of turbulent transport is used
to compute the vertical eddy viscosity and diffusivity in the three-dimensional
equations. The model contains six sediment transport processes—advection,
turbulent diffusion, settling/flocculation, deposition, erosion, and bed evolu-
tion. It is a "process-based" model rather than a "conceptual," "descriptive,"
or "phenomenological" model.
Limitations: The model may need long computation times, which results
in high computation costs. Detailed data are required for simulations and
calibration. The model may have low efficiency when it is applied to a mean-
dering river, as a rectangular domain is used in the model.
SED-2D.
General description: SED-2D is a finite element hydrodynamic, cohesive
sediment transport model for vertically averaged estuaries, rivers, and other
unstratified water bodies (Hayter 1987).
Capabilities and strengths: This model simulates two-dimensional surface
water flow and cohesive sediment transport. The effects of bottom, internal,
and surface shear stresses and the Coriolis force are represented in the equa-
tions of motion. The following sediment-related properties are calculated:
sediment bed structure (bed density and shear strength profiles, bed thickness
and elevation), net change in bed elevation over a given interval of time, net
vertical mass flux of sediment over an interval of time, average amount of
time sediment particles are in suspension, and the downward flux of sediment
onto the bed. It can be efficiently applied to water bodies having complex
geometries due to the employment of a finite element numerical scheme.
Limitations: SED-2D may not be suitable for long and continuous simula-
tion application due to computation costs. This model has not been com-
pletely tested.
•I c o
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
HEC-6.
General description: HEC-6 is designed to simulate one-dimensional,
steady, gradually varied water and sediment flow problems.
Capabilities and strength: The model can predict long-term trends of
scour and deposition in a stream channel. It can be used to predict reservoir
sedimentation, degradation of channel bed downstream from a dam, and the
influence of dredging activities. Local inflows and outflows of water and
sediment from tributaries and/or diversions can be included. It can analyze
channel contraction required to either maintain navigation depths or diminish
dredging requirements. Its strength is its ability to simulate hydraulic sorting
and bed armoring. This is done by sediment transport and scour/degradation
computations performed by grain-size fraction.
Limitations: HEC-6 is unable to directly simulate meandering phenome-
non, local scouring, bank erosion, and width adjustment. It is not suitable for
rapidly changing flow conditions. Equilibrium sediment transport capacity is
assumed. Density currents and bed forms are not accounted for.
RIVMOD.
General Description: RIVMOD is an unsteady, hydrodynamic and sedi-
ment transport riverine model that describes the longitudinal distributions of
flow and sediment concentration in a one-dimensional water body through
time (Hosseinipour and Martin 1991).
Capabilities and strength: The model allows prediction of gradually or
rapidly varying flows through time and space. It includes time-varying lateral
inflows. The sediment transport submodel predicts the transport of sediment
through the channel network and the scour/deposition processes as well as bed
level variations due to scour or deposition of materials. It can be applied to
noncohesive sediments (sand) and/or cohesive (fine) materials.
Limitations: Flows are assumed to be advectively dominant, and the
effect of eddy diffusivity is neglected. Water surface slope is assumed to be
small. The model in its present form does not include armoring and channel
stabilization. The cohesive sediment transport submodel does not account for
suspended sediment deposition and resuspension.
DYNHYD.
General description: DYNHYD is a simple link-node hydrodynamic
model that simulates variable tidal cycles, wind, and unsteady flows. It pro-
duces an output file that can be linked with the contaminant model WASP4
(described below) to supply the flows and volumes to the water quality model
(Ambrose et al. 1987).
1 64
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
Capabilities and strength: When linked to WASP4, it simulates the
movement and interaction of pollutants within the water. Driven by variable
upstream flows and downstream heads, simulations typically proceed at 1- to
5-min intervals. The resulting unsteady hydrodynamics are averaged over
large time intervals and stored for later use by the water quality program.
Limitations: No sediment transport simulations.
RMA(SED-2D).
General description: RMA (SED-2D) is developed for sediment problems
in rivers, lakes, and estuaries.
Capabilities and strength: This is a two-dimensional, unsteady model. It
can compute water surface elevations, current patterns, flow distributions
around islands, flow at bridges having one or more relieving openings, flow
in contracting and expanding reaches, flow into and off-channel storage for
hydropower plants, flow at river junctions, and general flow patterns. The
model can be used to compute sediment transport, deposition, and erosion in
two-dimensional open channel flows. It is applicable to clay and/or sand bed
sediments.
Limitations: Lengthy simulations are not feasible because of computation
costs. It is not designed for nearfield problems where flow structure interac-
tions are important. Variations in velocity or constituent concentration with
depth are not predicted. Only a single grain-size sediment can be analyzed,
and armoring is not addressed.
DAMBREAK.
General description: DAMBREAK is a dam-break flood forecasting
model. The model consists of a breach component, which utilizes simple
parameters to provide a temporal and geometrical description of the breach
(Fred 1988).
Capabilities and strength: This model computes the reservoir outflow
hydrograph resulting from the breach via a broad-crested weir flow approxi-
mation, which includes effects of submergence from downstream tailwater
depths and corrections for approach velocities. Also, the effects of storage
depletion and upstream inflows on the computed outflow hydrograph are
accounted for through storage routing within the reservoir.
Limitations: No sediment transport simulations.
1 65
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
CE-QUAL-RIV1.
General description: CE-QUAL-RTV1 is a one-dimensional (longitudi-
nal), water quality model for unsteady flows in rivers and streams. Output
from the hydrodynamic part is used to drive the water quality model (Environ-
mental Laboratory 1989).
Capabilities and strength: The model allows simulation of a branched
river system with multiple hydraulic control structures, such as run-of-the-
river dams, waterways, locks and dams, and regulation dams. The model was
developed to simulate highly unsteady flows that can occur on regulated
streams.
Limitations: No sediment transport simulations.
Contaminant transport models
Mass balance models for contaminants may be employed to estimate poten-
tial changes in contaminant concentrations for conditions prior to and after
remediation. The mass balance models can be used to predict chemical con-
centrations in various media (water, sediments, and fish). These estimated
concentrations can be used to calculate potential risks over time. The mass
balance models vary in their complexity, from simple analytical calculations
used to give rough screening level results to fully complex iterative models
that can predict the transport and fate of chemicals throughout time.
In the application of the contaminant exposure models, the rate of change
in mass (accumulation) is equated to the transport of a contaminant into, out
of, and within the system (via water flows or sediment flows for those mate-
rials that sorb to sediments), the mass added to the system (via point and
nonpoint loadings) minus the outputs and the quantities transformed and
degraded within the system (via processes such as volatilization, biodegra-
dation, and photodegradation). The output expected from the contaminant
exposure model includes estimated contaminant concentrations in water and
sediments (both paniculate and dissolved forms) as well as estimates of mass
fluxes due to inflows and loadings, outflows, degradation, and transformation
processes. Depending on the level of the modeling effort, the transport (via
water and sediments) may be described or predicted using hydrodynamic and
sediment transport models, which are then coupled with the contaminant
model. Table 13 suggests several contaminant transport and fate models that
could be utilized in a no-action alternative modeling study.
WASP4.
General description: WASP is a generalized modeling framework for
contaminant fate and transport in surface waters. Based on the flexible
1 fifi
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
Table 13
Suggested Fate and Transport Models
Fate and Transport Models for Organic Chemicals and Metals
Name
WASP4
EXAMS II
SMPTOX3
Source
USEPA, Athens
USEPA, Athens
USEPA, Athens
Dimension
1,2,3-D
1,2,3-D
1-D
Solution
Technique
Time Variable
Steady State
Analytical
Sediment
Transport
Y
N
N
Linked
w/Hydro.
Y
N
N
compartment modeling approach, WASP can be applied in one, two, or three
dimensions. WASP4 predicts dissolved and sorbed chemical concentrations in
the bed and overlying waters (Ambrose et al. 1987).
Capabilities and strength: This model is time variable and can simulate
three chemicals and three sediment size fractions simultaneously. The model
contains descriptive sediment resuspension/settling algorithms that allow for
the modeling of sediment transport. The model provides linkages to hydro-
dynamic models that provide changing flows and volumes to WASP on a
time-step-to-time-step fashion.
Limitations: The model does not have the kinetics for simulating metals
and oily wastes, although metals can be simulated descriptively using empiri-
cal distribution coefficients.
EXAMS II
General description: EXAMS is a generalized modeling framework based
on the WASP4 transport system for contaminant fate and transport in surface
waters. Based on the flexible compartment modeling approach, it can be
applied in one, two, or three dimensions. EXAMS predicts dissolved and
sorbed chemical concentrations in the bed and overlying waters (Burns and
Cline 1985).
Capabilities and strength: This model can run in a steady state or a
quasi-dynamic mode, three chemicals simultaneously. It is effective for doing
rapid evaluations of contaminant fate and transport. The model executes in
both an interactive and batch mode.
Limitations: The model is difficult to apply to a specific site. The model
does not simulate solids settling and resuspension.
Chapter 8 Contaminant Losses for the No-Action Alternative
167
-------
SMPTOX3.
General description: SMPTOX3 is a simplified analytical steady-state
model that can calculate the distribution of contaminants in water and
sediments.
Capabilities and strength: The model requires very few data to calculate
the distribution of chemicals. The model uses travel times and calculates total
chemical, sorbed chemical, and dissolved chemical. The model has interac-
tive data entry and graphical simulation results. The model can be used for
conducting screening-level calculations.
Limitations: The model is an analytical steady-state model with rudimen-
tary sediment/benthos algorithms.
Food chain models
A food chain model is a mass balance model for contaminants where the
rate of change in mass (accumulation) in each component of the food chain is
equated to the transport of a contaminant into and out of that component (via
ingestion, gill exchange, excretion, etc.) as well as internal changes that may
occur due to growth (dilution). The food chain model enables one to assess
the impact of remedial actions on contaminant concentrations within the food
chain, given variations in concentrations derived from the contaminant expo-
sure model. Outputs from a food chain model include time-varying estimates
of contaminant concentrations in each component of the food chain (Suarez et
al. 1986).
Summary
There are no fixed set of procedures for conducting No Action modeling
exercises. The approach that is taken is site specific and requires various
scenarios to be investigated and compared with the future exposure scenarios
projected for the site.
168
Chapter 8 Contaminant Losses for the No-Action Alternative
-------
9 Dredged Material Treatment
Dredged material that is contaminated to the extent that it requires
decontamination or detoxification in order to meet environmental cleanup
goals may be treated by one or more of a number of physical, chemical, or
biological treatment options. Treatment technologies reduce contamination
levels, contaminant mobility, or toxicity for the dredged material by one of
four ways:
a. Destroying the contaminants or converting the contaminants to less
toxic forms.
b. Separating or extracting the contaminants from the sediment solids.
c. Reducing the volume of contaminated material by separation of cleaner
sediment particles from particles with greater affinity for the
contaminants.
d. Physically and/or chemically stabilizing the contaminants in the
dredged material so that the contaminants are fixed to the solids and
are resistant to contaminant losses by leaching, erosion, volatilization,
bioaccumulation, or other environmental pathways.
Destruction technologies include incineration, vitrification, chemical treat-
ment, and biological treatment. Separation or extraction technologies include
solvent extraction, soil washing, and thermal desorption. Particle separation
technologies include hydrocyclones, classifiers, flotation, and screens. Stabili-
zation or immobilization technologies include a variety of solidification tech-
niques such as addition of lime and fly ash or addition of Portland cement to
create a solid product without free water. A comprehensive discussion of
process options for various treatment technologies is provided in Averett et al.
(1990). Guidance on the selection and implementation of sediment treatment
alternatives is available in the "ARCS Remediation Guidance Document"
(USEPA 1994a).
Other components are always involved for remediation alternatives that
involve treatment. Sediment is usually removed from the bottom of the water-
way by dredging, transported to the disposal site, and conditioned for treat-
ment and/or temporarily stored in a pretreatment facility prior to treatment.
159
Chapter 9 Dredged Material Treatment
-------
Treatment processes generate solid and liquid residue, as well as air emis-
sions. These streams may be subjected to further treatment or disposal.
Estimation of contaminant losses from the steps leading up to treatment and
from disposal have been discussed in previous chapters, and contaminant
losses from liquid effluents were discussed previously. Most treatment
processes include treatment of air emissions as an integral unit operation of
the process with the treated gas stream being released to the atmosphere. For
treatment processes, fractions of the contaminant in the feed to the process
may end up in the following compartments:
a. Fraction destroyed or detoxified within the treatment process.
b. Fraction remaining in the treated dredged material.
c. Fraction released to the atmosphere.
d. Fraction associated with dilute liquid effluents.
e. Fraction associated with concentrated liquid effluents.
/. Fraction associated with solids enriched with contaminants due to
separation of cleaner solids or adsorption media.
All treatment technologies do not generate all of the compartments listed
above.
Contaminant Loss Pathways From Sediment
Treatment Trains
Figure 39 illustrates the potential contaminant release points from sediment
treatment processes. Not all treatment processes generate all of the air emis-
sions and effluents shown in Figure 39. The components of the treatment
train for a specific type of technology will dictate which pathways or compart-
ments are important for that particular technology. The one component com-
mon to all treatment processes is the solids disposal block for the treated
sediment and other solid residuals. These materials will generally be sent to a
disposal site and are subject to the same pathways as disposal of untreated
sediment. However, the contaminant levels in the treated sediment will be
considerably reduced compared with untreated sediment, and the contaminant
loads from treated sediment in a disposal site would be expected to show a
corresponding reduction. Disposal pathway testing is recommended for the
treated sediment to estimate the magnitude of these releases. Concentrated
contaminant streams are usually transported to a hazardous waste treatment
facility for destruction or disposal. These facilities are presumed to have best
available treatment and state-of-the-art controls; therefore, contaminant losses
from this phase of the treatment operation will be assumed to be minimal.
' ' *-* Chapter 9 Dredged Material Treatment
-------
ADSORBENT/
SLUDGE/OIL
OFF-SITE
TREATMENT
FACILITY
DREDGED^
MATERIAL^
SEDIMENT
TREATMENT
SYSTEM
PARTICLE
SEPARATION,
FEEDING, ETC.
CONCENTRATED
CONTAMINANT
(OIL)
Figure 39. Contaminant losses from sediment treatment process trains
The applicability and importance of other emissions and effluents shown on
Figure 39 for several technologies are discussed in the paragraphs that follow.
Thermal destruction
The most common type of thermal destruction technology is incineration,
which has been demonstrated to be highly effective in destroying organic
contaminants in soils and sediments. The process basically involves heating
the sediment to temperatures ranging from 1200 to 2900 T1 in the presence
of oxygen to burn or oxidize the organic compounds in the sediment. Most
incinerators include a primary and a secondary combustion chamber. The
primary chamber partially destroys the contaminants and volatilizes the
remainder of the contaminants from the sediment, which are further oxidized
in the secondary chamber. The sediment's residue after incineration is a dry
ash or, for some innovative incineration processes, a dense slag or a glass-like
product. The gases from the combustion chamber pass through an emission
control system, which usually consists of a scrubber system, prior to release
to the atmosphere from the stack. The stack emissions are the contaminant
1 To obtain Celsius (C) temperature readings from Fahrenheit (F) readings, use the following
formula: C = (5/9) (F-32).
Chapter 9 Dredged Material Treatment
171
-------
loss pathway of most concern for incineration systems. Environmental regula-
tions may require destruction and removal efficiencies (ORE) of 99.9999 per-
cent for incineration of organic contaminants. The ORE is calculated as the
fraction of the contaminant mass fed to the incinerator that is released from
the stack. It usually does not include the residual contaminant in the treated
ash, nor does it include the wastewater and solids generated by the scrubber
system. Residual organics in the ash and the scrubber releases are expected to
be much less than 1 percent. However, volatile heavy metals, such as mer-
cury and lead, may be volatilized in the combustion chambers and be released
in the flue gas or concentrated in the scrubber wastewater. The wastewater
and ash may receive further treatment to remove the contaminants or reduce
contaminant mobility; whereas, the flue gas is released to the environment and
constitutes a contaminant loss.
Thermal desorption
Thermal desorption physically separates volatile and semivolatile contami-
nants from sediment by heating the sediment to temperatures ranging from
200 to 1000 °F, usually in an inert atmosphere. Water, organic contaminants,
and some volatile metals are evaporated from the sediment solids and are
subsequently captured or destroyed by the emission control system. Conden-
sation, scrubbing, adsorption, incineration, and paniculate control processes
are typical emission control system operations. Dust generated during the
drying process is captured by cyclones or paniculate filters. Potential contam-
inant loss streams include air from the emission control system stack, con-
densed oils and organic contaminants, condensed water or scrubber water,
collected dust, the residue after treatment, and contaminated activated carbon
or other adsorption media. Except for the stack gases, these streams will be
subjected to further treatment or disposal practices. The oil will be sent to a
hazardous waste treatment facility, the wastewater streams will be treated,
probably onsite, and the particulates and residues will likely be placed in a
disposal facility, either a landfill or a CDF. Since thermal desorption pro-
vides no treatment for heavy metals with the exception of mercury, metals in
the solid residue will be a potential source for leachate contamination in the
disposal site. Handling and transport of the dry, powdery residue will require
control measures to minimize losses of contaminants as dust.
Biological treatment
Biological treatment processes use microbes to degrade or transform
organic contaminants to less toxic or nontoxic compounds. Process options
for bioremediation of sediments include bioslurry reactors, land-treatment
systems, composting, and contained treatment facilities. Biosiurry systems
produce a treated residue, air emissions, and wastewater. The other types of
biotreatment systems generate a treated residue and may potentially generate
air emissions and leachate. Wastewater, or leachate collection and treatment,
and emission controls for bioslurry, contained land treatment, and composting
172
Chapter 9 Dredged Material Treatment
-------
systems would likely be part of the treatment train. Emission estimates for
land-treatment systems and for contained-treatment facilities without emission
controls could be made using techniques described for CDFs.
Extraction processes
Extraction processes are nondestructive processes that generally separate
contaminated sediment into solids, water, and an oily fraction containing the
contaminants extracted from the sediment. A number of different solvents
may be used for the extraction, including water with surfactants (soil wash-
ing), acetone, methanol, kerosene, triethylamine, and supercritical propane or
carbon dioxide. Most of the solvents are recovered and recycled into the
process. Potential contaminant losses for extraction processes are the waste-
water separated by the processes, the contaminant-rich oil, and the solids
residue. The wastewater would likely be treated onsite, and the oil phase
would be sent to a hazardous waste treatment and disposal facility. Most of
these processes can be closed to the atmosphere and do not have a positive gas
release to the atmosphere. Extraction processes for organics usually do not
affect heavy metals, and the metals remain with the residual solids. Removal
of heavy metals may be accomplished by a separate extraction train using an
acid or a chelating agent as the solvent. This train would require a concentra-
tion step for the heavy metals, which would have to be handled and disposed
as a contaminated material.
Chemical processes
Chemical processes are destructive processes that use reagents, tempera-
ture, or pressure to drive a chemical reaction with the contaminants convert-
ing them to environmentally acceptable materials. A major class of chemical
processes for sediment are the dechlorination processes for chlorinated hydro-
carbons such as PCBs. These processes generally operate in a closed environ-
ment with a minimal release to the atmosphere. Wastewater will be generated
by the process and will likely be treated onsite. The residual sediment will
contain traces of the organic contaminants and most of the heavy metals origi-
nally present in the sediment. Other chemical processes that involve gas-
phase reactions produce a stack emission that should be considered as a
contaminant loss stream.
Immobilization processes
Immobilization processes alter a sediment's physical and/or chemical char-
acteristics to reduce the potential for contaminants to be released from sedi-
ment when placed in a disposal site. Once placed in the disposal site, similar
techniques as used for confined disposal may be used to estimate losses, par-
ticularly by leacnates. Air emissions during the process of mixing sediment
with reagents or binders are likely, particularly when the solidification
173
Chapter 9 Dredged Material Treatment
-------
reaction generates heat, such as for Portland cement and pozzolan processes.
Several years ago, solidification was reported to destroy PCBs, but later
investigations proved that the PCBs were volatilized during processing rather
than being changed by the process. One of the advantages of solidification
processes is that they may virtually eliminate the effluent pathway and mini-
mize the leachate pathway, since free water in the sediment is usually
absorbed by the binder or becomes a part of a hydrated product.
Particle separation processes
Particle separation processes are usually considered as pretreatment
processes rather than treatment processes, particularly where the objective is
to remove oversized material from the sediments to avoid interference with
subsequent processing steps. However, they also offer treatment advantages
by separation of the clean fraction of a sediment from the more contaminated
sediment fraction in order to reduce the volume of material requiring more
costly treatment. Many commonly used options are available from the mining
and materials processing industries. Likely choices for sediments are hydro-
cyclones, screens, classifiers, and froth flotation. Most of these operations
process sediment as a slurry; therefore, a wastewater discharge or effluent will
be produced by the process. With or without treatment, contaminant will be
released with this effluent. Also, air emissions are possible due to the agita-
tion created by most of these processes. If volatilization proves to be an
important loss, the processing units may be housed and the emissions collected
and treated. Furthermore, the "clean" fraction will not be contaminant-free
and will represent a potential loss of organic and inorganic contaminants at the
disposal site. The contaminant-rich fraction will be subjected to other treat-
ment processes and the losses during these processes must be considered.
Techniques for Estimating Contaminant Losses
During Treatment
The wide range of chemical and physical characteristics for contaminated
sediment, the strong affinity of most contaminants for fine-grain sediment
particles, and limited application of treatment technologies to contaminated
sediment offer challenges to development of estimating or modeling techniques
to estimate contaminant losses for various contaminant and treatment tech-
nology combinations. Basic mathematical models are likely available for
simple process operations, such as extraction or thermal vaporization, applied
to single contaminants in relatively pure systems. However, such models
have not been validated for the sediment treatment technologies discussed here
because of the limited database for evaluation of treatment technologies for
contaminated sediment or soils and because of the wide range of sediment
physical and chemical characteristics that impact treatment processes. Devel-
opment of models for specific treatment technologies is beyond the scope of
this study.
1 74
Chapter 9 Dredged Material Treatment
-------
Standard engineering practice for evaluation of the effectiveness of treat-
ment technologies for any type of contaminated media (solids, liquids, or
gases) is to perform a treatability study for a sample that is representative of
the contaminated material. In a management review of the Superfund Pro-
gram, the U.S. Environmental Protection Agency (USEPA 1989) concluded
"To evaluate the application of treatment technologies to particular sites, it is
essential to conduct laboratory or pilot-scale tests on actual wastes from the
site, including, if needed and feasible, tests of actual operating units prior to
remedy selection." The performance data generated by treatability studies
will usually provide the contaminant concentrations for the residual sediment
following treatment. Contaminant concentrations and weights for side streams
generated by a technology can also be determined from treatability studies, but
the need for this information must be clearly identified as one of the objectives
of the treatability study so that appropriate data will be collected. Treatability
studies may be performed at bench-scale and/or pilot-scale level. Features of
each of these treatability study types are discussed below.
Bench-scale treatability studies
Bench-scale studies simulate the basic operation of a treatment process, but
are performed in a laboratory using a small volume (1 to 20 t) of sediment.
Individual operational parameters, such as chemical dosages, temperatures, or
retention times, and variable waste characteristics can be evaluated for a num-
ber of different conditions. Bench-scale tests generally use laboratory glass-
ware and carefully controlled conditions. The weights of solid or slurry and
liquid streams can be accurately measured, which can be coupled with con-
taminant concentrations for each stream to provide a mass balance around the
process for contaminants of concern. Side streams that include solid and
liquid phases should be separated and each phase quantified to provide infor-
mation needed to estimate the effectiveness of effluent treatment processes.
One of the limitations of bench-scale testing is that the volumes of side
streams generated may be too small for contaminant analysis at low concentra-
tions. Gaseous emissions are more difficult to collect and measure, and air
pollution control processes are more difficult to emulate in the laboratory in
conjunction with the solids treatment processes. Other limitations of bench-
scale studies include the volumes of the side streams produced may be insuffi-
cient to evaluate follow-on treatment technologies, and associated contaminant
losses for the side streams, and contaminant losses for pretreatment and mate-
rials handling processes are difficult to evaluate.
Pilot-scale treatability studies
Pilot-scale treatability studies are performed using significantly larger
volumes of sediment and using equipment that is similar to prototype process-
ing equipment but reduced in scale. Pilot tests are of sufficient scale to mini-
mize the physical and geometric effects of the test equipment on treatment
performance and simulate effects such as mixing, wall effects, generation of
175
Chapter 9 Dredged Material Treatment
-------
residues, heat transfer, or other factors in performance of the process. Infor-
mation on performance, design, and cost are much improved over bench-scale
tests. USEPA (1989) stated "Pilot-scale testing produces the most accurate
data on residuals generation, cross-media impacts, and treatment train require-
ments." Contaminant controls and losses can be evaluated for the primary
unit operation and for auxiliary unit operations used to control side streams
produced by the process, including gas streams and materials handling opera-
tions. Pilot studies can be planned to provide a mass balance for contaminants
of concern around the process train, thereby providing the information to
predict contaminant losses. Pilot studies are much more expensive to per-
form, and are generally executed after selection of a technology for a particu-
lar site based on technology screening and bench-scale testing.
Important contaminant loss components for treatability testing
Table 14 summarizes the important components of the treatment technolo-
gies discussed previously that should be evaluated during treatability study
testing in order to estimate contaminant losses. The ARCS program per-
formed this type of testing for a number of process options, and the reader is
referred to the reports for these tests for detailed information on the relative
magnitudes of each of the components for each type of technology. As was
stated earlier, these processes and treatability studies for these processes are
strongly influenced by sediment chemical and physical characteristics. Gener-
alization of the magnitude of these components into a table of guidance values
can be misleading without complete information on how the treatability study
was performed and complete laboratory data.
176
Chapter 9 Dredged Material Treatment
-------
n
a
a.
to
n
a.
Table 14
Important Contaminant Loss Components for Treatment Technologies
Contaminant Loss
Stream
Residual Solids
Wastewater
Oil/Organics
Leachate
Stack gas
Adsorption Media
Scrubber water
Particulates
(Filter/
cyclone)
Treatment Technology Type
Biological
X
X
Chemical
X
X
Extraction
X
X
X
X
Thermal
Desorption
X
X
X
X
X
X
Thermal
Destruction
X
X
X
X
Immobilization1
X
X
Particle
Separation
X
X
X
1 Immobilization is a special case for contaminant loss estimates in that its primary objective is to reduce leaching of contaminants from the sediment. Long-term
contaminant losses must be estimated using leaching tests and contaminant transport modeling similar to that used for sediment placed in a CDF. Leaching could be
important for residual solids for other processes as well.
-------
10 Example Application to
Contaminated Sediments in
the Buffalo River
Introduction
The remedial project discussed in this section is provided only for
discussion and illustration purposes—the loss calculations are "paper"
exercises. No actual field implementation is endorsed nor has occurred as
a consequence of this report.
This section describes example contaminant release calculations for a
selected area of concern. The calculations illustrate the types of site-specific
engineering assumptions that are required for implementation of the estimation
techniques described in previous parts of this report. Example calculations are
provided for losses from the following remediation components and
remediation alternatives: sediment removal (dredging), in situ capping (non-
removal remediation alternative), disposal without treatment in a CDF, and
treatment by thermal desorption.
Depending on the remediation component or alternative, various types of
results are obtained including concentrations, fluxes, and mass release rates.
In each case, however, the results are reduced to one common denominator-
contaminant mass loss per cubic meter of sediment remediated. Contaminant
loss estimates were normalized with respect to the volume of sediment for
remediation to facilitate comparison of losses among remediation components
and alternatives. To put loss estimates on a common basis, judgment is
needed about applicable time scales for analysis. Judgment about which con-
taminant loss mechanisms to include and a priori treatment process
effectiveness also affects loss comparisons.
Most of the calculations were implemented on commercially available
mathematical software (MATHCAD Version 4.0, Mathsoft, Inc., Cambridge,
MA) that allows the user to present equations as if they were written on engi-
neering paper. In one case, public domain software (the Hydrologic Evalua-
tion of Landfill Performance computer model) was used to estimate leachate
178
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
seepage from an upland CDF. Preparation of this report did not involve
computer model development, and no code was written to implement any of
the estimation techniques. Readers are directed to the fact that a single
computer code is not available for implementation of the various estimation
techniques described in this report. Commercially available mathematical
software is preferred by the authors over commercially available spreadsheet
type software because of problems with confirming if user-developed spread-
sheet algorithms are error free. With commercially available mathematical
software, the user is freed from the tedious task of checking cell addresses,
consistency of cell addressing, mysterious numbers in cells (unit conversion
factors), and the sequential logic behind extensive calculation suites.
Site Description
The Buffalo River area of concern was selected for this effort. This area is
shown in Figure 40. Two locations within the river were considered for
demonstration of calculating the contaminant losses from the hypothetical
implementation of remedial technologies, Dead Man's Creek and the Mobil
Oil area. Contamination in the Dead Man's Creek area includes PAHs in the
upper 50 to 100 cm of sediment. The volume of contaminated sediment at
this site is approximately 10,000 yd3. The Mobil Oil site includes about
40,000 yd3 of contaminated sediment, again in the upper 50 to 100 cm.
The sediments in both areas are composed of silts and clays with some
sands. Sediment samples from both sites also contain approximately 2 percent
organic carbon, which can sorb PAHs. A summary of sediment properties at
these sites is found in Figure 41.
Chemical analyses of the sediment from each site were used to identify
target contaminants to be used in the analysis and their concentrations. Four
PAHs (anthracene, benzoanthracene, benzopyrene, and phenanthrene) were
selected as indicator contaminants. Figure 41 lists two levels of contamina-
tion, an average and a high level. Average concentrations were determined by
simple averaging within the contaminated sediment regions. High concentra-
tion levels were determined by the average plus twice the standard deviation
among the samples. The high concentration levels would represent the
95-percent confidence limit if the sample concentrations were distributed
normally. The data, however, were not normally distributed with respect to
concentration. Although the data were not normally distributed, the high level
concentrations calculated were about the same as the highest observed concen-
trations. Because the concentrations at Dead Man's Creek were slightly
higher than those at the Mobil Oil site and the sediments were otherwise
similar, example contaminant loss release calculations focus on estimating
contaminant losses from the Dead Man's Creek site. All other conditions
being identical, the more concentrated sediment would be expected to result in
higher contaminant loss rates.
179
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Q.
CD
E
r
o
CD
6
CD
05
180
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
SITE 1: DEAD MAN'S CREEK
Volume: 10,000 cu yd
Sediment: silty clay, 40% porosity, bulk density 1.5 g/cm3, mean particle size 0.022
mm, 2% organic carbon
CONTAMINANT
Anthracene
Benzoanthracene
Benzopyrene
Phenanthrene
SITE 2: MOBIL OIL
LOW LEVEL (MEAN)
860 ^g/kg
11 50 //g/kg
770 //g/kg
1780 //g/kg
HIGH (MEAN + 2ff)
2990 //g/kg
4450 //g/kg
2 7 60//g/kg
5 930//g/kg
Volume: 40,000 cu yd
Sediment: silty clay, 43% porosity, bulk density 1.4 g/cm3, mean particle size 0.02
mm, 2.4% organic carbon
CONTAMINANT
Anthracene
Benzoanthracene
Benzopyrene
Phenanthrene
LOW LEVEL (MEAN)
800 //g/kg
540 //g/kg
340 //g/kg
1420 //g/kg
HIGH (MEAN + 2ff)
2200 //g/kg
1800 //g/kg
780 //g/kg
3790 //g/kg
Figure 41. Sediments and contaminants in Buffalo River AOC
In addition to the sediment and contaminant properties identified in Fig-
ure 41, river conditions influence contaminant losses during certain remedial
activities such as capping in place. The median discharge in the Buffalo River
has been estimated at 300 cfs (27.9 m3/sec). A rating curve has been
developed for various locations in the river. Dead Man's Creek is located
about 2.8 km from the lake discharge of the river and the rating curves
estimated at 2.4 km from the lake are1
Depth(m) h = 0.00258 Q + 7.2
(77)
Velocity(-) v =
sec
1
0.4089 + 1HI]
(78)
where Q is flow (m3/sec). These rating curves are valid for the region near
Dead Man's Creek for river discharges up to about six times the median flow.
The width of the river is about 82 m at this location. River conditions
described by Equations 77 and 78 were used to estimate losses associated with
in situ capping.
1 Personal Communication, 1992, U.S. Army Engineer District, Buffalo, Buffalo, NY.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
181
-------
Mean and high-sediment contamination levels for Dead Man's Creek were
selectively used in the contaminant-loss calculations that follow. Mean and
high-sediment contamination levels were used to estimate losses during dredg-
ing to illustrate the range in loss estimates that can be obtained using real site
characterization data as input. High sediment contamination levels were used
to estimate losses for in situ capping. Using the high sediment contamination
levels to estimate losses for in situ capping is a worst case scenario since
surficial sediments are more recent and often times cleaner. Mean sediment
contamination levels were used for estimating effluent and effluent treatment,
leachate, and volatile losses from pretreatment and disposal facilities because
dredging and dredged material placement/disposal tends to mix sediments.
Mean sediment contamination levels were also used to estimate losses for
thermal desorption processing of dredged material for the same reasons.
Comparison of the alternatives was based on losses calculated using mean
sediment contamination levels except for in situ capping.
Contaminant Losses During Dredging
Contaminant losses during dredging were estimated for both clamshell
(mechanical) dredging and cutterhead (hydraulic) dredging of the Dead Man's
Creek site.
Clamshell dredge
Calculations for contaminant losses during clamshell dredging are pre-
sented in Figures 42-45. Sediment parameters were either measured or
estimated from the available data. A key parameter in the evaluation of con-
taminant losses during dredging is settling velocity, which was estimated from
the mean grain size using Stoke's Law. This law is valid for dilute suspen-
sions of uniform grain-size particles and a Reynold's number less than 1
(negligible inertial effects). Given the range of grain sizes in typical
sediments, a measured settling velocity would be preferred.
A 10-yd3 open clamshell bucket was assumed. A closed bucket could be
used. Barnard (1978) estimated that closed buckets reduce turbidity by 30 to
70 percent compared with open buckets. Applicability of correction factors
based on Barnard (1978) to the Collins (1989) equations for sediment resus-
pension, however, has not been demonstrated. Advancements in closed
bucket technology that are currently available are not represented in either the
Barnard (1978) or Collins (1989) data. In a remediation project, the most
technologically advanced and cost-effective closed bucket would be preferred
over a conventional open bucket. The techniques presented and illustrated in
this report for conventional open buckets can be used to prepare loss estimates
for comparison with vendor-supplied information on currently available
closed-bucket technology.
1 B2
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Sediment parameters
Median grain diameter d - 0.022- mm '
Particle density p s = 2.65--=—
cm3
Water density pw = 1-551
cm3
Water viscosity \L - 1.3MO"2- poise
Settling velocity V o = d2-g-— V 3 = 0.001 • JL
18-ji sec
V3-d-pw
Reynold's Number N Re r N Re = ° °06
(Npe<1, required) fl
Clamshell parameters
Bucket volume V cb = 10-yd3
Characteristic length of bucket L cb = (2-V cb) L cb = 8.143 -ft
Clamshell cycle time \ cb = 120-sec
Vcb vd3
Dredged material production rate W = W = 300 •—
Tcb hr
V conj
Minimum dredging time = 33.333 -hr
V
conr
Resuspended sediment cone. Cp =0.0023-10" -pw-
(near bucket)
(Eq 10 of Text) C p = 555.S26-^
,-6 _ / L Cb
Figure 42. Clamshell dredge losses: sediment and dredge properties
1 83
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Clamshell Operation
Fraction of cycle in various modes (not used in this analysis)
Falling = 40%
Out of water = 10%
Rising = 40%
Dredging depth n b 2()'^
Bohlen sweep area correction y -4
Calculation of sediment release
Particle resuspension rate R p j-L cb2 C p R p = 695 43 •——
(Eq 11 of Text) t cb sec
Normalized resuspension rate -— = 10.915 • — -
Nakai (1978)"observed resuspension rate of 11.9 to 89 kg/m3 from bucket dredging
Figure 43. Clamshell dredge losses: resuspension calculations
184
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
4 contaminants
1-Anthracene
Avg Cone.
High Cone.
i = 1..4
.- H, =
kg '
kg
2- Benzoanthracene c
3- Benzopyrene
4- Phenanthrene
^ H =4450-10'6-^
kg 2 kg
770.]0-6.gm H 276(MO-6.gm
kg
kg
= 1780-
^ H = 5930-lQ-6.^
kg 4 kg
Mean release rate RM = R p-C & ' High release rate RH = R p-H.
Mean
Contaminant
Release
RM.
High
Contaminant
Release
RK
Contaminant
gm
hr
2.15
2.88
1.93
4.46
<— Anthracene
<- Benzoanthracene
<— Benzopyrene
<- Phenanthrene
Figure 44. Clamshell dredge losses: contaminant release
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
185
-------
Normalized contaminant loss in milligrams per cubic meter dredged.
N
M
N
H
"
-P
W
Normalized
Mean
Contaminant
Loss
High
Contaminant
Loss
N
Contaminant
H.
/mg\
U3/
mg
9.4
12.6
8.4
19.4
32.6
48.6
64.7
<- Anthracene
<- Benzoanthracene
<— Benzopyrene
<— Phenanthrene
Figure 45. Clamshell dredge losses: normalized contaminant loss
The characteristic length of the bucket (8.14 ft)1 was estimated by assum-
ing that the bucket was triangular in shape. The cycle time from collection of
sediment, raising the bucket, depositing the dredged material, and returning
the bucket to the riverbed was assumed to be 120 sec. Resuspended sediment
concentration near the bucket was estimated using Equation 10. Resuspended
sediment concentration in the water immediately surrounding the clamshell
bucket was estimated to be about 560 g/m3 or 560 mg/t. This concentration
would fall off rapidly with distance from the clamshell due to dilution.
The depth of dredging was assumed to be 20 ft, river depth near Dead
Man's Creek contaminated sediment area. The Bohlen sweep area correction
factor (typically 2-3) was chosen to be 2. The particle resuspension rate was
estimated (Equation 11) to be about 695 g/sec. Dividing the resuspension rate
by the dredge production rate provides an estimate of 10.9 kg of resuspended
sediment per cubic meter of sediment dredged. The resuspension estimate for
186
1 A table of factors for converting non-Si units of measurement to SI units is presented on
page xiii.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Dead Man's Creek is near the lower end of the range reported by Nakai
(1978) for mechanical dredges.
The contaminant release rate (Figure 43) was estimated as the product of
the sediment resuspension rate and contaminant concentrations in the sediment
(Equation 12). Contaminant release rates were estimated for both the mean
and high-level concentrations. The mean contaminant release rate ranged
from 1.9 g/hr of benzopyrene to 4.6 mg/hr of phenanthrene. The high con-
centration release rate ranged from 6.9 g/hr of benzopyrene to 14.8 mg/hr for
phenanthrene. Benzopyrene was the contaminant with the lowest concentra-
tion in Dead Man's Creek sediment, and phenanthrene was the contaminant
with the highest concentration in Dead Man's Creek sediment.
Dredging losses normalized with respect to the volume of sediment
dredged and were obtained as the product of normalized resuspension and
sediment contaminant concentrations (Figure 44). Normalized losses for mean
contamination levels were 8.4 mg/m3 to 19.4 mg/m3 for benzopyrene and
phenanthrene, respectively, and for high contamination levels, normalized
losses were 30 mg/m3 to 65 mg/m3 for benzopyrene and phenanthrene,
respectively.
Cutterhead dredge
Calculations for contaminant losses during cutterhead dredging are pre-
sented in Figures 46-49. A cutterhead dredge with a cutterhead measuring
2.5 ft long and 3 ft high was selected. The intake suction velocity, cutterhead
swing velocity, and cutterhead tangential velocity (rotational velocity) were
selected to be 0.625, 1.25, and 5 ft/sec, respectively. The fractional depth of
cut was selected to be 0.5. The production rate of the dredge of 371 yd3/hr
was estimated assuming that the suction velocity acted over the entire area of
the cutterhead and that the dredged material was 25-percent dry solids. Oper-
ational parameters are listed in Figure 46.
Estimation of the sediment resuspension rate followed that outlined in
Contaminant Losses During Dredging, specifically Equations 3 and 6. The
various coefficients (a, /3, FD,FF, etc.) were estimated and are shown in Fig-
ures 46 and 47. The estimated resuspended sediment concentration in the
vicinity of the cutterhead (Figure 47) was about 8 g/m3. This corresponds to
a resuspension rate of about 18 g/sec, or, normalized with the estimated pro-
duction rate, about 0.234 kg/m3. This is again in the lower range of the
resuspension rates from cutterhead dredges observed by Nakai (1978). Con-
taminant mass resuspension rates (Figure 48) were between 0.05 and 9.3 g/hr
for mean sediment contaminant concentrations and between 0.18 and 99 g/hr
for the high concentration sediment. Benzopyrene had the lowest release rate
of the contaminants examined and phenanthrene had the highest. Normalized
contaminant mass losses (Figure 49) were between 0.36 mg/m3 and 0.83 mg/
m3 for mean sediment contamination and between 1.3 mg/m3 and 2.8 mg/m3
for high-sediment contamination.
187
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Sediment parameters
Median grain diameter
Water density
Cutterhead parameters
Length of cutterhead
Height of cutterhead
Cutterhead characteristic size
Cutterhead size factors
Intake suction velocity
Ladder swing velocity
Blade velocity
Fractional depth of cut
d =0.022-mm
'w
gm
cm
Lch =2.5.ft
H
Lch-
ch
a =1.75
V: = 0.625-—
sec
V_ =1.25-
o
sec
_ft_
sec
Dp =0.5
D,
Dredged material production rate W - V s LCJ, D cjj
(assuming 25% solids)
Minimum dredging time
Dch = 3.557-ft
W = 185.255 «-
hr
W
= 53.98-hr
Figure 46. Cutterhead dredge losses: sediment and dredge parameters
Normalized mass loss estimates suggest that losses during cutterhead
dredging are less than 3 percent of the losses during clamshell dredging. In
general, contaminant release during cutterhead dredging is expected to be less
than during clamshell dredging.
188
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Calculation of sediment release
Empirical velocity significance factors: a =2.85 b =1.02
Burial coefficient
(Eq 5 of Text)
Other factors coefficient
(Eq 4 of Text)
Resuspended sediment cone.
(Near cutterhead)
(Eq 3 of Text)
Resuspension rate
D
0.41-(DF- I)7
= 1.472
FF.=
10
10
d-13.32
7.04
-2.05
Fp=0.09
rv-6
=pw-lor-.FFFD.
V;
V;
Cp =7.926-^
m
Rp =Cp-Vc.aHch.pLch Rp = 18.41--
R _ !,_
sec
Normalized resuspension rate —- = 0.468 •-
W
m
Nakai (1978) observed sediment releases between 0.1 and 45.2 kg/m3 for hydraulic
cutterhead dredges
Figure 47. Cutterhead dredge losses: resuspension calculations
Contaminant Losses During In Situ Capping
An alternative to dredging and treatment or disposal of contaminated sedi-
ment is capping in place with a clean layer of sediment. In situ capping iso-
lates contaminants from benthic organisms and the water column, significantly
reducing ecological impacts and allowing time for natural processes to remed-
iate contaminated sediment. In this example, times required for contaminants
to break through the cap, times to steady-state flux through the cap, steady-
state fluxes, and losses over the first 100 years normalized with respect to the
volume of sediment capped were estimated. The breakthrough time is the
time for the flux through the cap to reach 5 percent of the steady-state flux
while the steady-state time was arbitrarily selected as the time required to
reach 95 percent of the steady-state flux. The cap is assumed to be stable,
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
189
-------
i - 1 4
4 contaminants Avg Cone. High Cone.
1- Anthracene r _c
s, ~*
2- Benzoanthracene c - 1
560 lO-6-?? H, =2990-10-6-^
kg ' kg
ISO-IO-6-^ H ,4450.10-6-^
kg 2 kg
3- Benzopyrene =770106.^ H, =2760-l(r6.55
3 kg ^ kg
4- Phenanthrene c - ]
Mean release rate R ^ =
780 10' 6-^ H, =5930 10"6-^
kg 4 kg
RpCs High release rate RH -RpH.
V ; j V '
Mean High
Contaminant Contaminant
Release Release
RM,
Contaminant /^i
i \hrj
1 0.057
2 0.076
3 0.051
4 0.118
RH
/gm\
\hr /
0198 <- Anthracene
0.295 <- Benzoanthracene
0.183 <— Benzopyrene
0.393 <_ Phenanthrene
Figure 48. Cutterhead dredge losses: contaminant release
190
and contaminant transport through the cap is assumed to occur by diffusion,
retarded by sorption in the capping layer, and facilitated by natural organic
colloidal matter. Mass transfer processes driven by bioturbation were esti-
mated to be sufficiently fast that the capped zone populated by benthic animals
posed no effective mass transfer resistance. The benthic bioturbation mass
transfer coefficient and overlying water conditions are listed in Figure 50.
The calculations focus on diffusion-controlled losses. After the loss calcu-
lations for a cap with diffusion-controlled mass transfer are presented, loss
estimates are provided for advection-dominated mass transfer. The purpose of
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Normalized contaminant loss in milligrams per cubic meter dredged.
N
R,
M.
W
-•C,
N
H.
W
•H.
Mean
Contaminant
Loss
High
Contaminant
Loss
N
M.
N
H.
Contaminant
\m
\m
0.402
0.538
0.36
0.833
1.4
2.08
1.29
2.77
<- Anthracene
<- Benzoanthracene
<- Benzopyrene
<— Phenanthrene
Figure 49. Cutterhead dredge losses: normalized contaminant loss
the advection-dominated loss calculations is to compare diffusion-controlled
and advection-dominated losses and show that if diffusion controls, capping
can be a very effective remediation alternative.
Calculations were made for a cap with an effective depth of 50 cm. Effec-
tive depth is the actual depth of the cap minus the depth bioturbed by benthic
organisms. Properties of the cap (Figure 51) were assumed identical to the
properties of the underlying sediment (Dead Man's Creek). Although
contaminant-specific diffusivities are available or can be estimated, chemical
diffusivities in water do not vary widely and are all about 5 x 10"6 cm2/sec.
Diffusivities of the contaminants in water were, therefore, assumed to be 5 x
10~6 cm2/sec. In the cap, this diffusivity is modified by porosity and tortuos-
ity (Equation 48, Figure 51). Contaminant partitioning and reaction input
parameters are listed in Figure 52. The calculations were arranged to include
biodegradation by providing a characteristic reaction time, the compound half-
life. For the calculation summary shown in Figure 53, the compound half-
lives were assumed long enough (1 million to 100 million years) such that no
significant reaction occurred over the time of the calculations. Although these
half-lives may be too high to properly represent biodegradation, loss estimates
based on these half-lives will be conservative, that is, losses to the overlying
water column will be overestimated.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
191
-------
Water column properties (Assuming 2000 cfs flow)
Current speed
v = 0.002122-20000'8626-— v = 1 49-—
sec
sec
General contaminant/water properties
Diffusivity
Kinematic viscosity
(Water)
Schmidt Number
Dw=5-10
v =910-10
Sc = -
,-6 cm
sec
,-s cm2
sec
Sc = 1820
Volume/Depth of contaminated V ^^ .- 10000 yd
sediment
V,
Contaminated sediment area A =
cont
cont
Benthic bl m-t coefficient
(Turbulent boundary layer)
D
Ku = 0.036-
w
v-^/A
0.8
Figure 50. Contaminant losses for in situ capping: water/contaminant properties
Cap pore waters were assumed to contain natural organic colloidal material
at a uniform concentration of 25 mg/f (Figure 51). This colloidal material
can sorb contaminants, effectively increasing their "solubility" in the pore
water. This factor (1 + KocCdoc = 1 + KdCdoc/foc) was incorporated into the
estimation of the pore water concentration (Figure 54). For transient calcula-
tions, that is for the calculation of the breakthrough and steady-state times and
the transient flux-steady-state flux quotient, effective diffusivity was retarded
by sorption onto the immobile sediment phase. A retardation coefficient (R)
was defined that represents the total concentration of contaminant in the sys-
tem to the concentration in the water phase (Figure 53). This retardation
192
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Sediment/cap properties
Bulk density pj, =1.
Porosity e =0.4
Dissolved organic carbon
Bioturbed layer depth
Bioturbed layer diffusivity
Cap thickness .
(Total thickness - bioturbed depth)
Cap fraction organic carbon f^ =0.02
cm
doc
doc
bio
bio
liter
= 10-cm
= 10cm^
F
= 50-cm
Cap effective diffusion coefficient D eff = D w e
Deff=1.47-10
sec
Contamination Levels
4 contaminants
i =1..4
1-Anthracene
2- Benzoanthracene
3- Benzopyrene
4- Phenanthrene
Avg Cone.
High Cone.
„
sl
'•=? H, .=
kg 1
kg
•SZ H =4450-10 -^
kg ^ kg
'•^ R =
kg 3
kg
kg
kg
Figure 51. Contaminant losses for in situ capping: sediment/cap/contaminant properties
coefficient was adjusted by the factor 1 -f
transport by colloidal organic material.
to account for facilitated
Dissolved contaminant concentrations, which define the concentration
difference in the determination of the contaminant flux, was estimated by
assuming the pore water was in equilibrium with the sediment. If the sedi-
ment is above the critical loading, the predicted dissolved concentration could
exceed the solubility of the contaminant in water. The critical loading is the
sediment concentration at which equilibrium concentrations calculated using
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
193
-------
Contaminant Properties
4 contaminants Solubility Exchangeable Half-life1 rCjincap
1-Anthracene S, =0.045-^8 E,=1.0 T = 106.yr Kd=104-27^r.
liter ' l kg
-------
Retardation factor defined to account for colloidal transport
R = (epilson Rf )/(1+KocCdoc)
where Rf is as defined in Eq 57 of Text
cdoc = dissolved organic carbon conceentration
Koc = organic carbon partitioning coefficient
epilson is as defined in EQ 56 of Text —
Rf =
'doc
R
ER
oc
oc
Breakthrough time
(5% of steady flux)
Steady state time
(95% of steady flux)
= 0.54L
R
cap
tss =3.69.Lcap'
R
Fraction of compound remaining
after reaction at breakthrough
Results
Contaminant .
rxn.
(1= no reaction)
450
1373
1359
165
ss.
F
3073
9384
9285
1127
< Anthracene
< Benzoanthracene
< Benzopyrene
< Phenanthrene
Figure 53. Contaminant losses for in situ capping: calculation of tansient times
Contaminant loss estimates based on steady-state fluxes are unrealistic for
time frames significantly less than the time required to reach steady state.
The fraction of the steady-state flux occurring at times less than the steady-
state time was used to estimate fluxes over a time period of 100 years.
Equation 72 provides the transient flux-steady-state flux quotient as a function
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
195
-------
Overall mass tra
Note virtually all
to mass transfer
Calculated cont;
dissolved pore-v
cannot exceed !
Pore water cone
colloidally-sorbe
Steady-state flu:
Dissolved and p
and steady-stati
Contamin
T T -1
„ -, ^ cap 1 ^ bio 1
° Deff RDb,o Kb
T -1
of the resistance ., cm *"cap cm
. . ,1 "^ nv ~~ u.yz — — — u.yj —
is in the cap, e.g. ovi yr Deff yr
. / PbH [pbA
i R ' R
/ater concentration \ f ; \ f /.
solubility.
entration includes f K d
d contaminant. C pw = C w 1 + — C ^^
^OC
< through cap Flux ss - (K OV-C pwj
ore water (dissolved plus colloidal) concentrations
3 fluxes.
Cw. (Cpw.) Fluxss.
i V v i/ '
ant / UB \ / HB \ / m8 \
i \liter/ \liter/ \m2-yr/
1 8.01 11.74 0.108 < Anthracene
2 0.16 5.72 0.053 < Benzoanthracene
3 0.14 3.58 0.033 < Benzopyrene
4 56.09 63.45 0.581 < Rhenanthrene
Figure 54. Contaminant losses for in situ capping: steady-state flux— high concentrations
i QR
of time. Figure 55 shows the calculation setup for the first 100 years follow-
ing cap placement. Caution should be exercised when using Equation 72 as
indicated in Figure 55. The infinite series in Equation 72 is unstable for times
significantly less than the breakthrough time. This is indicated in two of the
four graphs in Figure 55. Anthracene and phenanthrene curves behave as
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Ratio of the flux at some time t to the steady-state flux is calculated
using equation 72. Define the quotient given in Eq 72 of the Text as the
Greek letter PHI, PHI = RA(t) / RA(t->infinity). Use 200 terms of the infinite
series and set up calculations for first 100 years.
A = 1 yr j =1 100 Note: i is the contaminant index and j is the year index.
t =Aj Examples of A'j and tj i^-20-yr t50 = 50-yr
200
-------
terms in the series is marginally effective when using single precision arithme-
tic. Roundoff error begins to degrade the results as the number of terms in
the series is increased. Actually, only 50 terms are needed to obtain con-
vergence for phenanthrene. Since breakthrough time is directly proportional
to the retardation factor, contaminants with low retardation factors may need
only few terms for the series to converge. For contaminants with high retard-
ation factors, the series is slow to converge.
Transient fluxes of anthracene and phenanthrene were integrated using the
trapezoidal rule to obtain the total emission per square meter for the first
100 years (Figure 56). Transient fluxes for benzoanthracene and benzopyrene
were not calculated because the transient flux was approximately zero for the
first 100 years. The results in mass per area are shown in Figure 56. These
results were then normalized with respect to the volume of contaminated
sediment capped (Figure 57). Normalized mass losses for in situ capping
were 1.3 x 10"8 and 0.05 mg/m2 for anthracene and phenanthrene, respect-
ively, and approximately zero for benzoanthracene and benzopyrene.
When an advective component is present, the above diffusional analysis of
contaminant losses for in situ capping can be seriously misleading. As previ-
ously discussed in Chapter 6, the significance of advection relative to diffusion
can be evaluated using the Peclet number (Equation 52). Figure 58 shows
anthracene breakthrough curves for Peclet numbers of 1, 10, and 50. Cap
thickness was used as the characteristic length.
Breakthrough curves were calculated using the Cleary and Adrian (1973)
finite length model for advection/dispersion with linear equilibrium-controlled
retardation. The same cap thickness (50 cm), same retardation coefficient for
anthracene (156, Figure 53), and same effective diffusion coefficient (Fig-
ure 51) used in the diffusional analysis were used to prepare the breakthrough
curves shown in Figure 58. The Peclet numbers represent three average pore
water velocities as follows: Pe = 1 and U = 10"7 cm/sec, Pe = 10 and U =
10"6 cm/sec, and Pe = 50 and U = 5 • 10'6 cm/sec. The instantaneous
advective flux is the product of average pore water velocity and contaminant
concentration at the cap-overlying water interface. Instantaneous fluxes at
Year 100 are shown in Table 15.
The instantaneous advective fluxes for Peclet numbers 1 and 10 are lower
and the instantaneous advective flux for Pe = 50 is larger than the steady-state
diffusional flux for anthracene shown in Figure 54. However, as shown in
Figure 53, the time to reach steady-state diffusional flux for anthracene is over
3,000 years. The times to breakthrough for advection, as defined by 5 per-
cent of the steady-state advective flux, are also shown in Table 15. Note that
the advective breakthrough occurs much more rapidly than for diffusion (Fig-
ure 58). In addition, the ultimate steady-state advective flux is U C0, or
identical to the advective flux without a cap. Thus, even a very small
advective flux can completely alter the contaminant loss picture for in situ
capping. In an advection-dominated system, the objective of capping is
1 98
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
The instantaneous flux at some time t is the product of the steady-state
flux (Table 29) and the flux ratio at time t (Table 30).
Instantaneous flux of contaminant i at time t is given by the following, where
j is the time index (1 to 100 years).
Flux, . ^
Flux. . ^
4,J
™
SS
Anthracene at 100 yr
s,'*i.j Example: Flux, ,00 = 1.32-10~3 -^-
m -yr
PhenathreneatlOOyr
J m -yr
Trapezoidal Rule:
j =2.. 100
IFlux. = Flux, -yr 4- - V (Flux. . 4- Flux. .Vyr <— IFlux is the integrated result.
i i, 1 *\ ^^^j \ '»J * '»J/
j
Anthracene
Benzoanthracene
IFlux =6.05' 10" «
1
Benzopyrene
m
Phenanthene
m
2.16-10~2 «5i
m
Figure 56. Contaminant losses for in situ capping: flux integration over time
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
199
-------
Time integrated results are multiplied by comtaminated area to obtain
total mass loss for the time period of integration. To normalize total
mass loss with respect to the volume, divide by the volume of
contaminated sediment.
IFlux.
N. =
d
<- Area / Volume = Depth
cont
Contaminant
Normalized 100 Year
Mass Losses For
Insitu Capping
N.
<— Anthracene
<— Benzoanthracene
<— Benzopyrene
<- Phenanthrene
Figure 57. Contaminant losses for in situ capping: normalized mass losses
200
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
600
Time (years)
Figure 58. Anthracene breakthrough curves for a 50-cm cap, r = 156, and
selected Peclet numbers
Table 1 5
Instantaneous Advective Anthracene Fluxes at Year 1 00 Through
50-cm Cap and Time Advective to Breakthrough (Based on 5 per-
cent of steady-state flux)
Pe
1
10
50
U
10'7
10'6
5 • 10'6
Flux
0.002
0.02
18
'•
= 235 years
= 1 00 years
= 33 years
Note: Pe: Peclet number, dimensionless.
U: average pore water velocity, cm/sec.
Flux: mg/m2«year.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
201
-------
containment until the contaminants are degraded or until a removal option can
be implemented.
Losses for Pretreatment/Confined Disposal
Effluent
Mechanical dredging and placement of dredged material in pretreatment
facilities for stockpiling and CDFs for disposal should result in minimal efflu-
ent losses since there is no conveyance water associated with mechanical
dredging. Effluent losses for mechanical dredging and placement, therefore,
are negligible. For mechanical dredging and hydraulic transfer to pretreat-
ment or CDFs, the losses will be similar to those discussed below for hydrau-
lic dredging and placement.
Effluent losses associated with hydraulic dredging and placement are best
estimated from column settling and modified elutriate tests. These data can be
applied to a specific facility design to predict losses or can be used in the
design phase to design a facility for target effluent quality. Column settling
and modified elutriate data are not available for materials from Dead Man's
Creek. Therefore, the a priori technique for estimating effluent quality
described in Contaminant Losses During Pretreatment was used to estimate
effluent losses. The a priori techniques involves Equation 22 and CEFs from
field studies to estimate effluent quality.
Palermo (1988) measured effluent quality and CEFs at five CDFs. The
five-site average CEF for metals was 0.986 (98.6 percent). Organic contami-
nants were not investigated except for PCBs at one site. The one-site CEF for
PCBs was 0.99 (99 percent). A CEF of 0.995 (99.5-percent containment)
was used to estimate effluent losses. A CEF value higher than the previously
measured CEFs is appropriate since the dredged material disposal operations
for which CEF data are available were maintenance dredging projects, not
remediation projects. It is assumed that remediation projects would put suffi-
cient emphasis on facility design and operation that containment performance
would be better than is typical for navigation maintenance projects.
Equation 22 in simple terms states that the fraction of contaminant mass
placed in a pretreatment or disposal facility lost during hydraulic filling is
1 - CEF. Thus, for a CEF of 0.995, the mass fraction lost is 0.005. An
estimate of mass loss was obtained by applying this factor to the sediment
contaminant concentrations and bulk density for Dead Man's Creek. Normal-
ized mass losses (product of contaminant concentration (mg/kg), bulk density
(kg/m3), and 0.005) are shown in Table 15. Sediment mean contaminant
concentrations (Figure 41) were used for these estimates because effluent from
hydraulic disposal operations tends to reflect the average dredged material
contamination levels. Normalized contaminant mass losses for hydraulic
202
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
placement of dredged material from Dead Man's Creek ranged from
5.8 mg/m3 for benzopyrene to 13 mg/m3 for phenanthrene.
The field CEFs on which effluent a priori loss estimates are based were
obtained using total (paniculate plus dissolved) effluent contaminant
concentrations. The effluent loss estimates in Figure 59, therefore, represent
paniculate and dissolved losses. Further, the a priori estimation technique for
effluent losses does not account for contaminant chemical properties. A priori
estimates are simply a faction of sediment contaminant concentrations, bulk
density, and the applied CEF. In spite of these limitations, effluent a priori
loss estimates are probably the most reliable a priori loss estimates that can be
made at this time.
Mechanical Disposal
Contaminant
Anthracene
Benzoanthracene
Benzopyrene
Phenanthrene
Hydraulic Disposal
Contaminant
Anthracene
Benzoanthracene
Benzopyrene
Phenanthrene
Normalized Mass Loss
No Treatment After Treatment
- Zero
~ Zero
~ Zero
~ Zero
Normalized Mass Loss
No Treatment
6.4
8.6
5.8
13
After Treatment*
1.5
2.0
1.3
3.0
* Carbon adsorption, 77 percent treatment effectiveness (Table 10)
Figure 59. Effluent losses for placement of dredged material from Dead Man's Creek,
Buffalo River
Effluent resulting from hydraulic placement of dredged material in pretreat-
ment and disposal facilities can be treated to reduce effluent losses and associ-
ated water quality impacts. Effluent could'be treated to reduce PAH losses.
Normalized PAH losses after treatment by carbon adsorption are also shown
in Figure 59. Normalized PAH losses after treatment were estimated by
applying the 77-percent removal efficiency listed in Table 10 for fluoranthene
by powdered activated carbon.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
203
-------
Leachate losses
Estimation of leachate losses requires estimation of leachate quality and
flow and site-specific information on pretreatment or CDF design. For this
example, upland pretreatment and CDFs were assumed.
Leachate flow was estimated using the HELP model in four simulations.
These simulations were conducted to estimate leachate flow from mechanically
placed dredged material in lined and unlined facilities and hydraulically placed
dredged material in lined and unlined facilities. A simple liner consisting of a
barrier soil (1-ft-thick) with a flexible membrane liner on top of the barrier
soil was simulated. The synthetic weather generator in HELP was used to
simulate climatological conditions for Buffalo, NY.
Spatial dimensions and dredged material properties for pretreatment and
CDF simulations were identical. Time frames for simulation were different.
The time for simulations of leachate flow from pretreatment facilities was
16 months, and the time for simulation of leachate flow from CDFs was
100 years.
Pretreatment and disposal facilities must be designed to handle dredge
production and, in the case of pretreatment facilities, meet the requirements of
the treatment process unit(s). For this example, a total processing time of
16 months for 10,000 yd3 was assumed for the treatment process unit(s).
Dredging could be scheduled in a variety of ways to satisfy this processing
rate. Since only 10,000 yd3 of material must be removed, the sediment could
be removed in a single dredging project requiring 3 to 5 days. For both
mechanical and hydraulic removal, it was assumed that all of the material
would be dredged and placed at one time in either a pretreatment or a disposal
facility.
Mechanical dredging and dredged material placement and hydraulic dredg-
ing and dredged material placement require different facility designs.
Mechanical dredging and placement involves minimal increase in volume over
the in situ volume of sediment. Hydraulic disposal, however, significantly
increases the volume over the in situ volume. A rule of thumb is that four
volumes of conveyance water becomes part of the dredged material for every
volume of in situ sediment. Thus, facility dimensions are affected by the
dredging method.
For mechanical disposal with negligible increase in dredged material vol-
ume over in situ sediment volume, a pretreatment or disposal facility must
hold 10,000 yd3 of material. Assuming an average depth of 6 ft, the facility
surface area is 45,000 ft2. The HELP model uses area to calculate total vol-
ume of seepage. General simulation parameters for facilities containing
mechanical placed dredged material are listed in Table 16.
A two-layer simulation for mechanical placement in an unlined facility was
conducted. The first or top layer is 6 ft of dredged material. The second or
204
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Table 16
Design Parameters for Leachate Flow From Unlined and Lined
Facilities Containing Mechanically Placed Dredged Material
Facility Design Parameters
Layer one - 6 ft, dredged material, vertical percolation layer
Layer two - Unlined Facility: 2 ft, foundation soil, vertical percolation layer.
Lined Facility: 1 ft, constructed barrier soil with FML.
Layer three - Lined Facility: 2 ft foundation soil, vertical percolation layer.
Soil and Dredged Material Properties
Porosity
Dredged Material = 0.40
Foundation Soil = 0.50
Constructed Barrier Soil = 0.4
Field capacity
Dredged Material = 0.32
Foundation Soils = 0.30
Barrier Soil = 0.32
Initial water content
Dredged Material = 0.40
Foundation Soil = 0.35
Barrier Soil = 0.35
Saturated hydraulic conductivity
Dredged Material = 1.0 E-06 cm/sec
Foundation Soils = 1.0 E-04 cm/sec
Barrier Soil = 1.0 E-07 cm/sec
Other
Evaporative zone depth = 12 in.
Type of vegetative cover - None
No runoff, all water must percolate or evaporate.
Area = 45,000 sq ft.
bottom layer is site foundation soil for which properties were assumed. When
a specific site is under consideration, soil properties from the site should be
used.
For hydraulic disposal with four volumes of water per volume of in situ
sediment, a pretreatment or disposal facility accepting all the material at once
must be able to store 50,000 yd3. For a storage volume of 50,000 yd3 and an
assumed depth of 8 ft, the surface area is 168,750 ft2. Hydraulically placed
dredged material was assumed to rapidly consolidate to a porosity of 0.75.
Further consolidation was not considered. Increasing the in situ sediment vol-
ume by the factor (0.75/0.4 = 1.875) and spreading this volume over
168,750 ft2 yields an estimated dredged material depth of 3 ft. The HELP
model simulations for hydraulic disposal were conducted as if the conveyance
water used to place dredged material in the facility were all discharged as
effluent, except for that retained in the dredged material. General simulation
parameters for facilities containing hydraulically placed dredged material are
listed in Table 17.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
205
-------
Table 17
Design Parameters for Leachate Flow From Unlined and Lined
Facilities Containing Hydraulically Placed Dredged Material
Facility Design Parameters
Layer one - 3 ft, dredged material, vertical percolation layer.
Layer two - Unlined: 2 ft, foundation soil, vertical percolation layer.
Lined: 1 ft, constructed barrier soil with FML.
Layer three - Lined Only: 2 ft foundation soil, vertical percolation layer.
Soil and Dredged Material Properties
Porosity
Dredged Material = 0.75
Foundation Soil = 0.50
Constructed Barrier Soil = 0.4
Field capacity
Dredged Material = 0.32
Foundation Soils = 0.30
Barrier Soil = 0.32
Initial water content
Dredged Material = 0.75
Foundation Soil = 0.35
Barrier Soil = 0.35
I Saturated hydraulic conductivity
Dredged Material = 1.0 E-06 cm/sec
Foundation Soils =1.0 E-04 cm/sec
Barrier Soil = 1.0 E-07 cm/sec
Other
• Evaporative zone depth = 12 in.
• Type of vegetative cover - None
• No runoff, all water must percolate or evaporate.
• Area = 168,750 sq ft
Table 18 lists total percolation from facilities containing mechanically and
hydraulically placed dredged material for 16-month and 100-year simulations.
These leachate flow estimates are for percolation from the foundation soil
layer. Although the percolation estimates were obtained using a vertical per-
colation simulation, leachate could move in all directions, including lateral
movement through the confining dikes.
The pore water contaminant concentrations (Figure 60) were estimated
using Equation 25-b. Equation 25-b includes the facilitated transport factor
(1 + KocCdoc). Leachate contaminant concentrations were assumed to remain
constant over time. For contaminants with high-distribution coefficients, such
as PAHs, this is a good assumption. For contaminants with low-distribution
coefficients, the assumption of constant-contaminant concentration overesti-
mates contaminant losses.
206
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Table 18
Totals for 16-Month and 100-Year HELP Model Vertical Percola-
tion Simulations
Placement
Mechanical
Hydraulic
Design
Unlined
Lined
Unlined
Lined
Total Percolation
1 6-Month (cu ft)
33,573 (0.31)
42 «0.1)
75,250 (0.69)
38 «0.1)
100-Year
(Thousand cu ft)
1,495 (2.2)
3.2 «0.1)
5,555 (8.2)
5.6(0.1)
Note: Numbers in parentheses are pore volumes of water displaced.
Figure 61 shows the sensitivity of contaminant concentration to the distri-
bution coefficient. Dimensionless time (horizontal axis in Figure 61) is the
number of pore volumes of water displaced. Large distribution coefficents
(> 100 f/kg) tend to keep contaminant concentrations low, but constant for
long times. Small distribution coefficients (< 10 f/kg) impose initially high
contaminant concentrations that rapidly decline.
Normalized mass losses for the leachate pathway were obtained as the
product of pore water contaminant concentration and total volume of leachate
divided by the in situ volume of sediment requiring pretreatment or confined
disposal. Normalized mass loss calculations for leachate are shown in
Tables 18 and Figure 60. Mean sediment concentrations were used to esti-
mate leachate losses because percolation tends to mix waters with varying
contaminant concentrations. The LV matrices in Figure 62 are transpositions
of the LV matrices (one column matrices) to one-row vectors, not the LV
matrix raised to the T power. Elements in the Nm and Nn matrices (Fig-
ure 62) are normalized mass losses by leaching. The column headings for
these matrices are the leachate volumes listed in Table 18, and the row
designations are the four contaminants listed in Figure 41.
Volatile losses
Sediment and contaminant characteristics for volatile loss calculations are
shown in Figure 63. Sediment characteristics include surface areas for
hydraulic and mechanical filling and sediment organic carbon content (foc).
Contaminant characteristics include mean sediment concentrations, distribution
coefficients, molecular weights, molar volumes, solubilities in water, vapor
pressures, dissolved water concentrations (estimated from sediment concentra-
tions and distribution coefficients), Henry constants, diffusivities in water,
overall liquid phase mass transfer coefficients (for an assumed wind speed of
15 mph), and gas-side mass transfer coefficients. Estimation of dissolved
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
207
-------
V =10000-ycr
Hg=10~ -gm
= 0.02
doc
-25-
mg
liter
Mean Leachate Contaminant Concentrations (Dead Man's Creek)
i =1..4
1-Anthracene
r -«rn
C „ -860
si kg
- r> iu ^ , ,™
2-Benzoanthracene C ,, -1150-
S
3-Benzopyrene
• 4-Phenanthrene
-
kg
kg
C = 1780 —
4 kg
pwi
,4.27 liter
kg
liter
"kg"
.-6.14 liter
=10
C3 kg
n3.73 liter
C4 " "kg"
Koc.'Cdoc
K f
oc. oc
1
Leachate Volumes From HELP Model
Facility Design Time Mechanical Hydraulic
Pretreat Unlined 16 mo LVm =33573-ft3 LVh =75250-ft3
CDF Unlined 100 yr LV = 1495000- ft3 LVu =5555000-ft3
u
Pretreat Lined
CDF Lined
16 mo
LV =42 ft3
LVu =38-ft3
100yr LVm =3200-ft3 LVh ^5600-ft3
m4 4
Figure 60. Contaminant losses by leaching: leachate concentrations and volumes
water concentrations did not include contaminant mass associated with colloi-
dal organic matter because to volatilize from water, contaminants must be
truly dissolved. Various constants, such as temperature, viscosity of water,
atmospheric pressure, and molar volume of air, are also assigned values in
Figure 63.
Calculations of volatile emission rates from ponded water are shown in
Figure 64. The basic volatile flux equation for ponded water (Equation 32)
was modified to an emission equation by multiplying flux by the ponded water
208
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
1.00
0.75
0.50
O
0.25
0 200 400 600
T (PORE VOLUMES ELUTED)
BULK DENSITY - 1 kg / 1
POROSITY - 0.50
PORE WATER VELOCITY - 1 x 10"5 cm / sec
DISPERSION COEFFICIENT - 1 X 10"5 cm2/sec
LENGTH - 100 cm
800
1000
Figure 61. Fraction initial contaminant concentration remaining in leachate
for various distribution coefficients
surface area for hydraulic filling. The surface area for hydraulic filling (A2)
as previously noted in the calculation of leachate losses is larger than the
surface area for mechanical filling. In the example calculations shown in
Figure 64, background air quality was assumed to be clean, that is, PAH con-
centrations in the background air were assumed to be negligible. Figure 64
also shows the calculation of normalized mass loss by volatile emission from
ponded water for anthracene, benzoanthracene, benzopyrene, and phenan-
threne. These calculations are applicable to pretreatment and disposal facili-
ties because the ponded water holding time is about the same. Operation of
each type of facility requires holding water long enough for adequate solids
settling. For facilities of similar size, as assumed in these calculations,
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
209
-------
V =10000-ycr
foe =0.02
doc
=25-
mg
liter
Mean Leachate Contaminant Concentrations (Dead Man's Creek)
i =1..4
1-Anthracene
2-Benzoanthracene
3-Benzopyrene
4-Phenanthrene
= 86oHf KOC ,104-27.^
i kg
_ If)6.i4 liter
°°2 kg
kg
II o
f\ r^O
kg
'kg
K
OC
^
=1780-^ K™
kg °°4
3 kg
..3.73 liter
kg
Krc foc
UC. \J^f
i
Facility Design
Pretreat Unlined
CDF Unlined
Pretreat Lined
i
Leachate Volumes From HELP Model
Time Mechanical Hydraulic
16 mo LVm .= 33573-ft3 LVh =75250-ft3
1 i i
100 yr LV = 1495000-ft3 LV u = 5555000-ft3
16 mo
LVu :=38-
CDF Lined 100 yr LVm =3200-ft3 LVh =5600-ft3
4 4
Figure 62. Contaminant losses by leaching: normalized mass losses
holding time requirements are similar. A 7-day holding time was used as
previously discussed in the calculation of leachate losses. Normalized mass
losses by volatilization from ponded water were highest for phenanthrene and
lowest for benzopyrene. These estimates represent maximum potential losses
210
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
V = 10000-y(T
Ug=10~6-gm
Loc
= 0.02
Aj = 45000 ft <— surface area of facility for mechanical filling
A2 - 168750 ft2 <- Surface area of facility for hydraulic filling
is
1.139 <- viscosity of water (centipoise) T = 288 <- temperature (K)
Ma =28.97 <-molecular wt of air
Va =20.1 <-molar volume of air (cc/mole)
P = 1 <— pressure (atm)
a
Mean Sediment PAH Concentrations
4 PAHs Avg Cone.
1-Anthracene
M, = 860--
1 kg
Distribution Coefficients Mol Wt
•, ,_,«4.27i. liter
2-Benzoanthracene M, = 1150-
3-Benzopyrene
4-Phenanthrene
kg
M, = 770-—
3 kg
M, = 178Q-HI
4 kg
r6.uf liter
0 ^V
6.of liter
-foc—
Mb =178.24
M b = 228.3
Mb =352.3
Mb =178.24
4
PAH Molar Volumes (Miller as cited by Mackay, Shiu, and Ma 1992) (cc/mole)
Vb =197 Vb =248 Vb =263 Vb :=199
\ 2, j 4
PAH Solubilities in Water (from Mackay, Shiu, and Ma 1992) (mg/L)
S, =0.075 S2 =0.014
=0.004
S4:=1.2
PAH Vapor Pressures (from Mackay, Shiu, and Ma 1992) (Pascals)
a = .09357 <— factor for converting Pascals to mm Hg
Pa =0.00141-a
i
Pa =4.1-10' -a
Pa . = 7.32-10"7-a
Pa =0.0161-a
Figure 63. Contaminant losses by volatilization: sediment and contaminant characteristics
(Sheet 1 of 3)
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
211
-------
Assume equilibrium for estimation of dissolved concentrations in ponded water.
i =1..4
M.
C. =
K
d.
C =
0.002
4.166-10
-5
3.85-10
0.017
-5
liter
Henry Constants (result is dimensionless)
H = 16.04-
Pa.'Mb.
T-S.
0.017
3.484-10"
3.36-10"4
0.012
PAH Diffusivity in Water <- result in cmA2/sec
D
13.26-Iff
A2.
. 1.14 /,, \0.589
i. • Vb;
D
A2
5.089-10
4.444-10"
4.293-10"
5.059-10"
Overall Liquid Side Mass Transfer Coefficient <- results in cm/nr
V x = 15 <- assumed wind speed in mph
= 19.6-V™-(DM\3-™
-• x \ ./ hr
KOL~
2.432
2.222
2.172
2.423
cm
> . __
hr
Figure 63. (Sheet 2 of 3)
212
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Gas-Side Mass Transfer Coefficients
Kg =3000-
18 cm
Mb hr
Kg=
953.356
842.373
678.112
953.356
cm
>
hr
PAH Diffusivities in air <- results in cm/sec
M
ri " Ma-Mb.
D
|M,-103-T1'75 2
' ' cm
Al.
2 sec
•P
D
Al
0.055
0.049
0.047
0.055
cm
t
sec
Figure 63. (Sheet 3 of 3)
because dissolved concentrations in ponded water were assumed to be equilib-
rium controlled and constant.
Calculations for volatile emissions from exposed dredged material solids
for mechanical and hydraulic filling are shown in Figures 65 and 66, respec-
tively. Calculations for mechanically filled and hydraulically filled facilities
use the same basic equation (Equation 39). However, values for total poros-
ity, air-filled porosity, and bulk density are different. The calculations
involve piecewise integration of Equation 39 over time using the Romberg
algorithm in MATHCAD. Piecewise integration was used to improve the
accuracy of the results. The flux equation was integrated over 16 months and
100 years to simulate exposure times for pretreatment and disposal facilities,
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
213
-------
Ponded Water Volatile Emission Rates - Hydraulic Filling
Ponded water volatile emission rates from mechanically
placed dredged material are assumed to be negligible.
EP=
2.114-10'
348.291
314.572
1.546-105
.mg
day
Assume 4 day retention time for adequate solids settling plus 3 days for
drawdown to yield total of 7 days emission time for ponded water.
Normalized mass losses by volatile emission from ponded water.
N,
7-day-E.
19.351
0.319
0.288
141.544
m
<- Applicable to both
pretreatment and
disposal facilities.
<- Anthracene
<— Benzoanthracene
I <— Benzopyrene
<- Phenanthrene
Figure 64. Volatile emission rates from ponded water—applicable to hydraulically filled
pretreatment and disposal facilities
respectively. As previously discussed for leachate losses, the pretreatment
facility will be needed for about 16 months, after which it will be closed. A
disposal facility, however, is permanently maintained. A 100-year simulation
time for a disposal facility was an arbitrary selection, influenced by
uncertainty about applicability of the basic flux equation for long-term
simulations.
214
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
For Mechanical Filling
e =0.4
Ea =0.1
<- Total porosity
<- Air filled porosity
gm
b = 1.5-2^- <-Bulk density of Dredged Material
cm3
PAH Diffusivity in Soil Gas
D
DAl.'ea
A3.
D
A3
1.605-10
1.422-10
-4
1.356-10
1.598-10
. 4
-4
cm
sec
Time-Integrated Flux From Mechanically Filled Dredged Material
MATHCAD's Romberg Integration, Tolerance set at 0.000001
TiFLUX = time integrated flux
Piecewise Integration Is Implemented Over Four Time Domains:
I: 0-1 day
II: 1 - 30 days
III: 30 days to 16 months
IV: 16 months to 100 years
TiFLUXJ. =
1-day
M.-H.
n-t
dt
0-day
TiFLUXJ j = 0.123 •—* <-Anthracene volatile flux from exposed sediment
m integrated over 0-1 day — Mechanical Filling
Figure 65. Volatile emission from exposed dredged material—mechanical filling (Sheet 1
of 4)
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
215
-------
TiFLUXJI. ~-
30-day
M.-H.
1-day
D
A3.
n-t 1
-\
dt
TiFLUXJIj = 0.179 -—5 <-Anthracene volatile flux from exposed sediment
m integrated over 1 to 30 days — Mechanical Filling
TiFLUX III. =
16-30-day
M.-H.
K
d.
30-day
jt-t
D
AS:
dt
TiFLUX JIIj = 0.658 •— ^-Anthracene volatile flux from exposed sediment
m2 integrated over 30 days to 16 months -
Mechanical filling
Figure 65. (Sheet 2 of 4)
216
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
TiFLUXJV. =
100-365-day
M.-H.
-d.
16-30-day
n-i
g;
dt
TiFLUXJVj =6.791 -^ <-Anthracene volatile flux from exposed sediment
m2 integrated over 16 months to 100 years -
Mechanical filling
TiFLUXJ 6mo. = TiFLUXJ. + TiFLUXJI. + TiFLUXJII.
TiFLUX lOOyr. -TiFLUX I. + TiFLUX II. + TiFLUX III. + TiFLUX IV.
— j| — j ~~1 "~1 1
TiFLUXJ 6mOj = 0.961 •— <- Anthracene volatile flux from exposed sediment
m2 integrated over 16 months - Mechanical Filling.
TiFLUX JOOyi-j = 7.752 •— <- Anthracene volatile flux from exposed sediment
m2 integrated over 100 years - Mechanical Filling.
Figure 65. (Sheet 3 of 4)
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
217
-------
Volatile Losses Normalized With Respect To the Insitu Volume of Sediment Dredged
Pretreatment Facility - Exposure Time = 16 months
NormVolLoss_Pt
TiFLUXJGmo.-A,
V
NormVolLossJ'l M =
0.525
0.001
0.001
1.579
m
<-Anthracene
<— Benzoanthracene
I <- Benzopyrene
<- Phenanthrene
Disposal Facility - Exposure Time Infinite - Losses Estimated For 1st 100 Years
NormVolLoss_Dis
TiFLUXJOOyr.-A,
NonnVolLossDis =
4.239
0.012
0.009
11.897
.mg
<—Anthracene
<— Benzoanthracene
m
3 <— Benzopyrene
<- Phenanthrene
Figure 65. (Sheet 4 of 4)
218
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Volatile Fluxes From Exposed Dredged Material (mud)
Hydraulic Filling
E =0.75
Ea:=0.2
<- Total porosity
<- Air filled porosity
b = 0.86-5^ <-Bulk density of Dredged Material
cm
PAH Diffusivity in Soil Air
D
10
3
Al.'Ea
D
A3.
4.602*10
4.076-10
-4
-4
3.886-10
4.58MO
-4
Flux From Hydraulically Filled Dredged Material
MATHCAD's Romberg Integration, Tolerance set at 0.000001
i =1..4
Volatile Flux Integrated Over 0 -1 day
rl day
TiFLUX I. =
M.-H.
0-day
0.133
71- 1
D
AS:
dt
TiFLUX I =
.-4
1.54-10
1.132-10
0.588
-4
<— Anthracene
<— Benzoanthrancene
mg
m2 <— Benzopyrene
<- Phenanthrene
cm
sec
Figure 66. Volatile emissions from exposed dredged material —hydraulic filling (Sheet 1
of 4)
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
219
-------
Volatile Flux Integrated Over 1-30 days
r 30-day
TiFLUXJI. =
M.-H.
TiFLUX II =
1 day
0.23
6.642-10~
4.918-10"
0.641
7t-t 1
-\
dt
<— Anthracene
<- Benzoanthrancene
tmg
m2 <— Benzopyrene
<- Phenanthrene
Volatile Flux Integrated Over 30 day -16 mo
i-16 30- day
TiFLUX 111. =
M.-H.
30-day
dt
rc-t
K
g,
TiFLUX 111 =
0.845
0.002
0.002
2.352
<— Anthracene
mg <- Benzoanthrancene
m2 <— Benzopyrene
<- Phenanthrene
Figure 66. (Sheet 2 of 4)
220
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Volatile Flux Integrated Over 16 mo -100 vrs
TiFLUX IV. =
— \
100365-day
M.-H.
-d.
30 16-day
n-1
dt
TiFLUX IV =
8.714
0.025
0.019
24.263
<- Anthracene
mg <- Benzoanthrancene
m2 <— Benzopyrene
<- Phenanthrene
Summations of Piecewise Integrations for 16 Months and 100 Years
TiFLUX 16mo. = TiFLUX I. + TiFLUX II. + TiFLUX III.
— i — i — i — i
TiFLUXJOOyr. = TiFLUXJ6mo. +- TiFLUXJV.
TiFLUX 16mo =
TiFLUXJOOyr =
1.208
0.003
0.002
3.58
9.922
0.029
0.021
27.843
<— Anthracene
^mg <- Benzoanthrancene
m2 <- Benzopyrene
<— Phenanthrene
<— Anthracene
tmg <- Benzoanthrancene
m2 <- Benzopyrene
<- Phenanthrene
Figure 66. (Sheet 3 of 4)
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
221
-------
Volatile Losses Normalized With Respect To the Insitu Volume Of Sediment Dredged.
Pretreatment Facility - Exposure Time =16 months
NormVolLoss Pt^
TiFLUX_16mo.-A2
NormVolLoss_Pt =
2.477
0.007
0.005
7.341
.mg
<—Anthracene
<— Benzoanthracene
m
3 <— Benzopyrene
<- Phenanthrene
Disposal Facility - Exposure Time Infinite - Losses Estimated For 1st 100 Years
NormVolLoss Disi
TiFLUX_100yr.-A2
NormVolLossDis =
20.346
0.059
0.044
57.092
<—Anthracene
<- Benzoanthracene
.m§
3 <- Benzopyrene
m
<— Phenanthrene
Figure 66. (Sheet 4 of 4)
222
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Figures 64 and 65 conclude with calculations of volatile losses normalized
with respect to the in situ volume of sediment dredged. Comparison of nor-
malized volatile losses in Figures 64 and 65 showed that losses were higher
for hydraulically filled facilities than for mechanically filled facilities. This is
due primarily to the larger surface area of hydraulically filled versus mechani-
cally filled facilities. Losses for pretreatment facilities were substantially
lower than losses for disposal facilities due to the lower exposure time for
pretreatment facilities.
Contaminant Losses During Treatment By Thermal
Desorption
Thermal desorption is one possible treatment option for removal of PAH
compounds from Buffalo River sediment. An ARCS pilot study of thermal
desorption treatment of Buffalo River sediment was performed in 1991, and a
report has been prepared describing results of this study (USAGE, Buffalo
District 1993). In order to evaluate the effectiveness of thermal desorption
and to collect design and operational data for future work, a monitoring pro-
gram was implemented. The monitoring program included all streams enter-
ing and exiting the thermal desorption system. These data provide a basis for
estimating mass of contaminant in each process stream, and, therefore, an
estimate of contaminant losses.
A process flow diagram for the Buffalo River pilot thermal desorption
(TD) unit is shown in Figure 67. Dredged material was screened prior to
feeding the thermal desorption unit to remove oversize material. In this case,
the oversize material consisted primarily of roots and debris. After screening,
the sediment was stored in covered 208-f, plastic-lined steel drums. Differ-
ences in contaminant concentrations before and after screening were not sig-
nificant, suggesting that losses during screening were minimal. Major outputs
from the thermal processor were the product solids, solids from a series of
cyclones that removed particulates in the air stream, condensed liquids from
the air stream, and a gas release from the stack. The system was designed to
collect two separate liquid streams, one an oil residue small in volume and
high in contaminant concentrations and the other a water stream high in
volume and low in contaminant concentrations. During the Buffalo River
Demonstration, these streams were difficult to separate and were similar in
contaminant concentrations. A full-scale TD unit would require additional
treatment of these liquid streams. Prior to release from the stack, the gas
stream passed through an activated carbon bed. Spent carbon from a full-
scale unit would require further treatment or disposal. The cyclone solids had
PAH concentrations on the same order of magnitude as the dredged material,
but the volume of cyclone solids collected was small relative to the volume of
dredged material treated.
Nine separate runs were evaluated in the pilot demonstration. However,
complete data sets are not available for every run. In particular, a limited
223
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
SPENT
CARBON
OFF-SITE
TREATMENT
FACILITY
CONCENTRATED
.CONTAMINANT
(OIL)
Figure 67. Process flow diagram for thermal desorption unit used in Buffalo River pilot
demonstration
number of quality-ensured air analyses are available. A single run having a
relatively complete data set was selected for estimating contaminant losses.
This run was labeled as A2 and was conducted on 23 October 1991. Operat-
ing data for this run included a retention time of 60 min in the thermal proces-
sor and a soil exit temperature of 480 °F. Mass balance data are presented in
Table 19. The mass of solids fed to the processor and exit streams were con-
verted to pounds per hour dry solids since most analyses were reported on a
dry weight basis. Table 20 provides the contaminant concentrations for each
stream. The concentration and mass flow rate were multiplied to yield a con-
taminant mass emission per hour. Finally, Table 20 normalizes the mass of
contaminant in each stream to the contaminant mass in the feed. This makes
it convenient to extrapolate the results to a site-specific feed with different
contaminant concentrations as shown in Table 21.
Table 21 also shows normalized contaminant concentrations in feed and
process streams. The normalized contaminant concentrations in Table 21
represent contaminant mass in each stream per cubic meter of in situ sediment
to be remediated. Most of the process streams could receive further treatment
or could be placed in a secure facility with negligible contaminant losses. The
results in Table 21 show that the cyclone catch represents the largest fraction
224
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Table 19
Buffalo River Thermal Desorption Pilot Study, Mass Balance Data
Stream
Feed
Treated solids
Cyclone catch
Condensate
Stack gas
Mass Rate
Total
Ib/hr
502
238
25
222
NA
Fraction
Dry
Solids
0.545
0.998
0.862
0.0095
NA
Mass Rate
Dry solids
Ib/hr
273.59
237.52
21.55
2.11
Table 20
Analysis of Buffalo River Thermal Desorption Pilot Study Data
Stream
Anthracene
Benzo(a)pyrene
Benzo(a)-
anthracene
Phenanthrene
Concentration, ng/g (dry weight basis)
Feed
Treated solids
Cyclone catch
Condensate
Stack gas
133
5
101
22.3
NA
545
10
260
21.1
NA
542
5
228
13.8
NA
670
37
618.
111
NA
Contaminant Mass Flux, mg/hr
Feed
Treated solids
Cyclone catch
Condensate
Stack gas
16.5447
0.5403
0.9629
2.2476
0.047
67.7958
1.0805
2.4788
2.1266
0.020
67.4226
0.5403
2.1738
1.3909
0.012
83.3453
3.9979
5.8920
11.1875
0.49
Fraction of Contaminant in Stream Compared with Mass in Feed
Feed
Treated solids
Cyclone catch
Condensate
Stack gas
Estimated carbon
load
1.0
0.0327
0.0582
0.1359
0.0028
0.7704
1.0
0.0159
0.0366
0.0314
0.0003
0.9158
1.0
0.0080
0.0322
0.0206
0.0002
0.9390
1.0
0.0480
0.0707
0.1342
0.0059
0.7412
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
225
-------
Table 21
Extrapolation of Pilot Study Data to Contaminant Loss Example
Problem
Fraction of Contaminant in Stream Compared With Mass in Feed
Feed
Treated solids
Cyclone catch
Condensate
Stack gas
Estimated carbon
load
1.0
0.0327
0.0582
0.1359
0.0028
0.7704
1.0
0.0159
0.0366
0.0314
0.0003
0.9158
1.0
0.0080
0.0322
0.0206
0.0002
0.9390
1.0
0.0480
0.0707
0.1342
0.0059
0.7412
Contaminant Concentrations in Example Sediment From Dead Man's Creek
Contaminant
concentration in
feed, ng/g
Anthracene
860
Benzoanthracene
1,150
Benzopyrene
770
Phenanthrene
1,780
Normalized Mass Concentration
Contaminant
concentration in
feed, mg/m3
Treated solids,
mg/m3
Cyclone catch,
mg/rn3
Condensate,
mg/m3
Stack gas, mg/m3
Estimated carbon
load, mg/m3
1,290
42.2
75.1
175.3
3.6
993.8
1,725
27.4
63.1
54.2
0.5
1,579.8
1,155
9.2
37.2
23.8
0.2
1,084.5
2,670
128.2
188.8
358.3
15.8
1,979
of contaminant in the residues requiring further treatment or disposal. The
one process stream that would be difficult to further control is the stack gas
(air emissions after carbon adsorption, Figure 67). The normalized concentra-
tions in this process stream, therefore, are normalized contaminant losses for
thermal desorption treatment, assuming other process residues receive further
treatment or disposal without contaminant loss.
226
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Comparison of Contaminant Losses
Overall
Figure 68 shows normalized PAH mass losses for five remedial alterna-
tives. These alternatives are listed in Table 22. Alternative I involves
mechanical dredging and mechanical disposal in an upland CDF. Controls for
m
E
120
100
80
60
40
20
co
§ •
CO
1
fi"
25
20
15
10
5
0
1 [ I
ANTHRACENE
50
40
30
20
10
II III IV
I I 1 I 1
BENZOPYRENE
I II III IV
LEGEND
^ WITHOUT CONTROLS
H WITH CONTROLS
500
400
300
200
100
ALTERNATIVE
I I I I I
BENZOANTHRACENE
I II
I I I
PHENANTHRENE
IV
Figure 68. Normalized PAH mass losses
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
227
-------
Table 22
Alternatives Considered for Remediation of Dead Man's Creek
Alternative
1
II
III
IV
V
Description
Mechanical dredging and mechanical disposal
in an upland CDF
Hydraulic dredging and disposal in an upland
CDF
Mechanical dredging and mechanical place-
ment in a pretreatment facility (equalization)
and thermal desorption processing of dredged
material solids
Hydraulic dredging and placement in a pre-
treatment facility (equalization and dewater-
ing) and thermal desorption processing of
dredged material solids
In situ capping
Controls
Liner
Effluent treatment by
carbon adsorption and
liner
Liner for pretreatment
facility
Carbon adsorption treat-
ment of pretreatment
effluent and liner for pre-
treatment facility
Assumes cap stability and
isolation from bioturbation
Alternative I are limited to lining the CDF to minimize leachate losses. Efflu-
ent controls are not needed for Alternative I since dredging and disposal are
mechanical. Alternative II involves hydraulic dredging and disposal in an
upland CDF. Controls for Alternative II include effluent treatment by carbon
adsorption and lining the CDF. Alternatives III and IV involve mechanical
and hydraulic dredging, respectively, and include pretreatment (equalization
and dewatering) and thermal desorption processing of sediment solids. Con-
trols for Alternative III are limited to lining the pretreatment facility to mini-
mize leachate losses. Effluent controls are not needed since dredging and
placement are mechanical. Controls for Alternative IV include treatment of
effluent from the pretreatment facility by carbon adsorption and lining the
pretreatment facility to minimize leachate losses. Alternative V is in situ cap-
ping and does not involve dredging.
As indicated in Figure 68, contaminant loss calculations showed that in situ
capping is superior to all other alternatives in terms of minimizing PAH
releases over a 100 year period. Diffusion-controlled PAH release associated
with in situ capping is estimated to be 1,000 to greater than 100,000 times
less than the next best alternative. Alternatives are ranked in order of
decreasing contaminant loss for each PAH in Table 23. Alternative IV with
controls is second best in minimizing losses for all four PAHs. Alternative II
without controls releases more PAH than any of the other alternatives, and
Alternative I without controls was next worst. Between the second best and
second worst alternatives, the relative order of the rankings vary with PAH.
In general, the rankings for anthracene and phenanthrene are similar, and the
rankings for benzoanthracene and benzopyrene are similar. Differences
between rankings for the anthracene-phenanthrene pair and the
228
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
Table 23
Alternative Ranking by PAH
PAH
Anthracene
Benzoanthracene
Benzopyrene
Phenanthrene
Ranking
V
IV with controls
III with controls
1 with controls
IV without controls
III without controls
II with controls
1 without controls
II without controls
V
II with controls
IV with controls
IV without controls
1 with controls
III with controls
III without controls
1 without controls
II without controls
V
II with controls
IV with controls
IV without controls
1 with controls
III with controls
III without controls
1 without controls
II without controls
V
IV with controls
1 with controls
III with controls
III without controls
IV without controls
II with controls
1 without controls
II without controls
Normalized Mass Loss,
mg/m3
1.32E-08
7.98
13.51
13.67
13.80
13.93
22.27
32.33
96.70
0
2.62
3.03
10.04
12.63
13.10
13.28
20.80
39.59
0
1.72
1.87
6.64
8. .42
8.61
8.73
13.96
26,80
0.05
26.98
31.52
36.78
39.12
42.21
61.3
135.3
457.9
benzoanthracene-benzopyrene pair are related to differences in chemical prop-
erties, as discussed below.
Alternative I
Figure 69 shows that most of the anthracene and phenanthrene losses for
Alternative I without controls were through the leachate pathway. Dredging
losses were second in relative significance for anthracene and phenanthrene,
and volatile losses were third in relative significance for these two chemicals.
Most of the benzoanthracene and benzopyrene losses were associated with
dredging, and leachate losses made up the rest of the losses for these two
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
229
-------
ALTERNATIVE I: WITHOUT CONTROLS
CLAMSHELL DREDGING W/CDF DISPOSAL
(mg/m3)
ANTHRACENE: 32.33
BENZOANTHRACENE: 20.80
BENZOPYRENE: 13.96
PHENANTHRENE; 135.3
Figure 69. Alternative I without controls
chemicals. Volatilization was insignificant for benzoanthracene and
benzopyrene.
Figure 70 shows losses for Alternative I with leachate controls (a lined
CDF). Dredging losses dominate losses for all four PAHs, especially benzo-
anthracene and benzopyrene. Lining a CDF, of course, does not increase
dredging losses. Because leaching losses have been significantly reduced,
230
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
ALTERNATIVE I: WITH CONTROLS
CLAMSHELL DREDGING W/CDF DISPOSAL
(mg/m3)
LEACHATE
ANTHRACENE: 13.67
BENZOANTHRACENE: 12.63
LEACHATE
BENZOPYRENE: 8.42
PHENANTHRENE: 31.52
Figure 70. Alternative I with controls
dredging losses represent a proportionally larger share of the total loss esti-
mate. Volatile losses are second in relative significance for anthracene and
phenanthrene, and with leachate controls in effect for Alternative I, leachate
losses of these two chemicals were of minor significance. Volatile losses
were negligible for benzoanthracene and benzopyrene, and leachate losses
were of extremely minor significance for these two chemicals.
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
231
-------
PAH losses for Alternative I were reduced by 39 (benzoanthracene) to
77 (phenanthrene) percent by lining the CDF. Controls were less effective for
benzoanthracene and benzopyrene than for anthracene and phenanthrene. The
differences in control effectiveness is due to differences in the significance of
dredging losses. For benzoanthracene and benzopyrene, dredging losses
comprised a greater share of the total losses in the without-controls alternative
than dredging losses for anthracene and phenanthrene. Therefore, imple-
menting leachate controls (no impact on dredging losses) has less effect on
benzoanthracene and benzopyrene losses.
The differences in primary loss pathways for different PAHs under Alter-
native I are related to differences in chemical properties. Anthracene and
phenanthrene are more mobile than benzoanthracene and benzopyrene. Solu-
bilities are higher (Figure 52), Henry constants are higher (Figure 63), and
distribution coefficients (Figure 52) are lower for anthracene and phenanthrene
than for benzoanthracene and benzopyrene. Thus, anthracene and phenan-
threne were lost through pathways involving large masses of water (e.g.,
leachate) and volatilization. Benzoanthracene and benzopyrene were lost
through pathways involving large masses of solids (dredging). Although the
solubilities of anthracene, benzoanthracene, benzopyrene, and phenathracene
are not high relative to many other chemicals, and distribution coefficients for
these chemicals are not low relative to many other chemicals, leachate losses
of these chemicals were significant for the unlined CDF option. Volatile
losses were significant for anthracene and phenanthrene and insignificant for
benzoanthracene and benzopyrene.
Alternative II
Figure 71 shows that most of the PAH losses for Alternative II without
controls were through the leachate pathway. Volatile losses were second in
relative significance for anthracene and phenanthrene, and effluent losses were
third in relative significance for these two chemicals. For benzoanthracene
and benzopyrene, effluent losses were second in relative significance, and
dredging losses were third in relative significance for these two chemicals.
Volatilization was insignificant for benzoanthracene and benzopyrene.
Figure 72 shows losses for Alternative II with controls (effluent treatment
and a lined CDF). Most of the anthracene and phenanthrene were lost
through volatilization. Effluent losses were second in relative significance for
anthracene and phenanthrene, dredging losses were third in relative signifi-
cance, and leachate losses were fourth in relative significance for these two
chemicals with effluent and leachate controls. Effluent losses dominate losses
for benzoanthracene and benzopyrene. Dredging losses were second in rela-
tive significance for benzoanthracene and benzopyrene, volatile losses were
third in relative significance, and leachate losses were fourth in relative signif-
icance for these two chemicals with effluent and leachate controls.
232
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
ALTERNATIVE H: WITHOUT CONTROLS
CUTTERHEAD DREDGING W/CDF DISPOSAL
(mg/m3)
DREDGING
.EACI
DREDGING
VOLATILE
ANTHRACENE: 96.70
BENZOANTHRACENE: 39.59
DREDGING
EFFLUENT
LEACHATE
BENZOPYRENE: 26.80
PHENANTHRENE: 457.9
Figure 71. Alternative II without controls
PAH losses for Alternative II were reduced by 77 (anthracene) to 94 (ben-
zopyrene) percent by implementing controls (Table 22). Losses were pri-
marily reduced by restricting leachate flow. Controls were more effective for
benzoanthracene and benzopyrene than for anthracene and phenanthrene. The
differences in control effectiveness was due to differences in the relative sig-
nificance of leachate losses. Benzoanthracene and benzopyrene leachate losses
comprised a greater share of the total losses in the without controls-alternative
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
233
-------
ALTERNATIVE H: WITH CONTROLS
CUTTERHEAD DREDGING W/CDF DISPOSAL
(mg/m3)
LEACHATE
DREDGING
ANTHRACENE: 22.27
VOLATILE
LEACHATE
BENZOPYRENE: 1.72
VOLATILE
LEACHATE
BENZOANTHRACENE: 2.62
LEACHATE
DREDGING
PHENANTHRENE: 61.3
Figure 72. Alternative II with controls
than leachate losses for benzoanthracene and benzopyrene. Thus, leachate
controls had more impact on those PAHs whose losses in the uncontrolled
alternative were primarily leachate losses.
The relative significance of loss pathways for Alternative II with controls
varied. For anthracene and phenanthrene, volatile and effluent pathways were
234
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
the primary loss pathways and volatile losses dominated. For benzo-
anthracene and benzopyrene, effluent and dredging pathways were the major
pathways and effluent losses dominated. The differences in the significance of
effluent and dredging losses is related to differences in the significance of the
volatilization pathway. For Alternative II with controls, volatilization was the
dominant loss pathway for anthracene and phenanthrene, but relatively insig-
nificant for benzoanthracene and benzopyrene. The Henry constants are
higher for anthracene and phenanthrene than for benzoanthracene and benzo-
pyrene; hence, a greater tendency for anthracene and phenanthrene to be lost
by volatilization. Relative to volatilization, dredging losses of anthracene and
phenanthrene were small for Alternative II with controls. Since volatilization
was insignificant for benzoanthracene and benzopyrene, dredging and effluent
losses accounted for a larger portion of the total losses of these compounds for
Alternative II with controls.
Alternative III
Figure 73 shows that PAH losses for Alternative III without controls were
primarily associated with solids losses during dredging. Stack gas losses from
the thermal desorption unit were second in relative significance for anthracene
and phenanthrene, and leachate and volatile losses were relatively minor for
these two chemicals. For benzoanthracene and benzopyrene, leachate losses
and losses in the stack gas from the thermal desorption unit were relatively
minor. Volatilization was negligible for benzoanthracene and benzopyrene.
Figure 74 shows the distribution of losses for Alternative III with controls
(lined pretreatment facility). Figures 73 and 74 are similar because lining the
pretreatment facility minimally reduces total losses and does not significantly
alter the distribution of losses for Alternative III.
Anthracene and phenanthrene losses from the thermal desorption unit were
significant relative to other losses, where as benzoanthracene and benzopyrene
losses from the thermal desorption unit were insignificant relative to other
losses. These differences can be explained on the basis of chemical proper-
ties. Anthracene and phenanthrene as previously discussed are more mobile
and tend to sorb less than benzoanthracene and benzopyrene. Thus, sorption
and retention in the carbon column treating stack gases from the thermal
desorption unit were greater for benzoanthracene and benzopyrene than for
anthracene and phenanthrene.
Alternative IV
Figure 75 shows that anthracene and phenanthrene losses for Alternative
IV without controls were distributed among treatment, leachate, effluent,
volatile, and dredging losses. Dredging losses were the least significant of all
the losses for these chemicals. Effluent was the major, though not the domi-
nant loss pathway, for anthracene. Losses from the thermal desorption unit
235
Chapter 10 Example Application to Contaminated Sediments/Buffalo River *-^^
-------
ALTERNATIVE III: WITHOUT CONTROLS
CLAMSHELL DREDGING W/PT AND TD
(mg/m3)
LEACHATE
VOLATILE
LEACHATE
ANTHRACENE: 13.93
BENZOANTHRACENE: 13.28
LEACHATE
LEACHATE
BENZOPYRENE: 8.73
PHENANTHRENE: 39.12
Figure 73. Alternative III without controls
were the major losses, though not the dominant loss, for phenanthrene.
Benzoanthracene and benzopyrene were lost primarily through the effluent
pathway.
Figure 76 shows the distribution of PAH losses for Alternative IV with
controls (effluent treatment and lined pretreatment facility). Thermal desorp-
tion losses and volatile losses from the pretreatment facility were the most
236
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
ALTERNATIVE III: WITH CONTROLS
CLAMSHELL DREDGING W/PT AND TD
(mg/m3)
LEACHATE
VOLATILE
LEACHATE
ANTHRACENE: 13.51
BENZOANTHRACENE: 13.10
LEACHATE
VOLATILE
BENZOPYRENE: 8.61
PHENANTHRENE: 36.78
Figure 74. Alternative III with controls
significant losses for anthracene and phenanthrene. Effluent and dredging
losses were the most significant losses for benzoanthracene and benzopyrene.
PAH losses for Alternative IV were reduced by 36 (phenanthrene) to
75 (benzoanthracene and benzopyrene) percent by implementing controls.
Most of the reduction in losses were due to effluent treatment. Controls were
more effective for benzoanthracene and benzopyrene than for anthracene and
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
237
-------
ALTERNATIVE IV: WITHOUT CONTROLS
CUTTERHEAD DREDGING W/PT AND TD
(mg/m3)
DREDGING
DREDGING
ANTHRACENE: 13.80
LEACHATE •
BENZOANTHRACENE: 10.04
DREDGING
LEACHATE
DREDGING
BENZOPYRENE: 6.64
PHENANTHRENE: 42.21
Figure 75. Alternative IV without controls
phenanthrene. The differences in control effectiveness was due to differences
in the relative significance of effluent losses. Benzoanthracene and benzopy-
rene effluent losses comprised a greater share of the total losses in the without
controls-alternative than effluent losses for anthracene and phenanthrene.
Thus, effluent controls had more impact on those PAHs whose losses in the
uncontrolled alternative were primarily effluent losses.
238
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
ALTERNATIVE IV: WITH CONTROLS
CUTTERHEAD DREDGING W/PT AND TD
(mg/m3)
ANTHRACENE: 7.98
LEACHATE
BENZOANTHRACENE: 3.03
LEACHATE
• LEACHATE
BENZOPYRENE: 1.87
DREDGING
PHENANTHRENE: 26.98
Figure 76. Alternative IV with controls
Effluent was a significant loss pathway for benzoanthracene and benzopy-
rene because other loss pathways such as volatilization and stack gas emission
from the thermal desorption unit comprised relatively minor shares of the total
losses. Anthracene and phenanthrene volatile losses from the pretreatment
facility and stack gas emissions from the thermal desorption unit were more
significant than effluent after treatment. Figure 76 suggests that further engi-
neering controls could be chemical specific. For example, a cover for the
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
239
-------
pretreatment facility could reduce anthracene and phenanthrene losses, but
would have little effect on benzoanthracene and benzopyrene losses. Addi-
tional effluent treatment would reduce losses of all four PAHs, but would
have a larger effect on the total losses for benzoanthracene and benzopyrene
than on the total losses of anthracene and phenanthrene.
Summary
Several insights are offered by the example calculations. These insights
are as follows:
a. Mechanical dredging followed by mechanical disposal in a CDF and
hydraulic dredging followed by disposal in a CDF results in approxi-
mately the same PAH losses when leachate and volatile losses are
neglected. In the former, PAHs are lost primarily at the point of
dredging. In the latter, PAHs are lost in the effluent. The total mass
loss is about the same. However, engineering controls are more prac-
tical for effluent than for dredging.
b. When leachate and volatile losses are considered, mechanical dredging
and mechanical disposal in a CDF result in lower PAH losses than
hydraulic dredging and disposal in a CDF. The primary difference is
in leachate losses. Leachate losses are higher for the hydraulic dredg-
ing option because hydraulic dredging adds water to the sediment that
is not removed during sedimentation as effluent. This water bulks the
sediment and, depending on site-specific foundation conditions and
hydrologic factors, may drain by gravity into underlying soils.
Leachate controls are, therefore, more likely to be cost-effective for
the hydraulic dredging and disposal option than for the mechanical
dredging and disposal option.
c. The significance of volatile losses is highly chemical dependent. Four
PAHs were included in the example calculations. For two, anthracene
and phenanthrene, volatile losses were significant for some alternatives.
For the other two, benzoanthracene and benzopyrene, volatile losses
were negligible or minor for every remediation option considered.
d. Certain conventional wisdoms about dredging and dredged material
disposal may need to be revisited. It is often said that many contami-
nants strongly sorb to sediment solids and, therefore, are not mobile.
The example calculations suggest that this may be somewhat mislead-
ing. No matter how large the distribution coefficient, reversible sorp-
tion implies a capacity and potential for desorption.
240
Chapter 10 Example Application to Contaminated Sediments/Buffalo River
-------
11 Summary and
Recommendations
Conclusions
General
The primary objective of this report was presentation of techniques for
estimating contaminant losses associated with various sediment remediation
alternatives. Preproject estimation of contaminant losses conducted early in
the planning process can indicate the relative merit of various remediation
alternatives. Intuitively, the alternative that minimizes contaminant losses is
the most environmentally protective alternative. Although risk assessment,
economics, feasibility, and other factors must be considered to fully evaluate
alternatives, preliminary screening or ranking of alternatives according to
estimated contaminant losses has merit because it is contaminant loadings
(losses) to the environment that result in exposure concentrations that impair
environmental quality. In addition, contaminant loss estimates provide some
of the input data needed to conduct risk assessments for remediation
alternatives.
Many environmental regulatory agencies are beginning to emphasize
assessment of total mass losses of contaminants in their evaluations of dredged
material management alternatives. Existing procedures such as the Corps of
Engineers Management Strategy (Francingues et al. 1985), the Dredged Mate-
rial Alternative Selection Strategy (Cullinane et al. 1986), the General
Decision-Making Framework (Lee et al. 1991), and the Interagency Technical
Framework for Evaluating Environmental Effects of Dredged Material Man-
agement Alternatives (USEPA/USACE) use analyses of contaminant migration
pathways to estimate environmental effects (for example, water column and
benthic toxicities). Estimated effects are compared with criteria established by
regulatory authorities to arrive at decisions regarding the suitability of an
alternative, including the need for restrictions. When acceptable combinations
of restrictions are difficult to identify, the existing procedures provide little
guidance for objectively evaluating tradeoffs between alternatives, including
the no-action alternative. The approach to comprehensive analysis of contami-
nant losses described in this report provides an objective, comparative
241
Chapter 11 Summary and Recommendations
-------
assessment methodology that engineers, scientists, planners, decision-makers,
and others involved in evaluation of sediment remediation alternatives may
find helpful.
Techniques are available for estimating contaminant losses associated with
most sediment remediation components and contaminant migration pathways
within remediation components. In some cases, a priori estimation techniques
are available that do not require data other than sediment characterization data
and other minimal project data. Pathway specific laboratory tests are avail-
able for some contaminant migration pathways that provide more reliable
estimates of losses than the a priori techniques. A priori techniques are suita-
ble for planning-level assessments. Techniques that use pathway-specific
laboratory data provide the type of loss estimates often called for by regula-
tory agencies that evaluate proposed remediation projects.
Availability and relative reliability of contaminant loss estimation tech-
niques are shown in Table 24. The state of development of predictive tech-
niques for estimating contaminant losses from remediation components varies
with the component and the loss pathway. For some remediation components,
there are no pathway-specific tests available. In these cases, a priori tech-
niques may be the only techniques available; however, a priori techniques are
not always available for all pathways of all components. The confidence and
accuracy of contaminant loss estimates depend on the state of development and
the amount of field-verification data available.
Table 24
Availability and Relative Reliability of Contaminant Loss Estimation
Techniques
Component or
Alternative
In Situ Capping
Open-Water
Disposal/Capping
Dredging
Transportation
Confinement
Treatment
No Action
Available
Yes
Yes
Yes
No
Yes
Yes
Yes
Reliability
Moderate
Moderate
Low
-
Variable
High
High
Ease of Use
Difficult
Difficult
Moderately Difficult
--
Moderately Difficult
Simple
Very Difficult
This report illustrates how overall pooled estimates for all pathways and
remediation components can be used to compare sediment remediation alterna-
tives in terms of effectiveness. Most of the available estimation techniques,
however, are not sufficiently developed or field verified to warrant decision
making on the basis of contaminant loss estimates alone. Even if the
242
Chapter 11 Summary and Recommendations
-------
estimation techniques were fully developed and field verified, it would not be
prudent to use estimated contaminant losses alone for decision making.
Using the a priori techniques described and illustrated in this report, it is
possible to identify major contaminant loss pathways for various alternatives.
This information can then be used to identify needs for laboratory testing to
provide sediment-specific parameters for refined estimates of contaminant
losses. For example, if for alternative A, pathways i and j are shown to be
relatively insignificant and pathway k is shown to be significant relative to
pathways i and j, then laboratory determination of sediment-specific parame-
ters for pathway k is indicated if the comparison of alternative A to other
alternatives is to be refined. Thus, the a priori techniques described in this
report can be used to allocate resources toward refining contaminant estimates
and, hence, evaluation of remediation alternatives.
This report includes a set of example calculations. Most of the calculations
were implemented on commercially available mathematical software that
allows the user to present equations as if they were written on engineering
paper. In one case, public domain software (the Hydrologic Evaluation of
Landfill Performance computer model) was used to estimate leachate seepage
from upland pretreatment and CDFs. Preparation of this report did not
involve computer model development, and no code was written to implement
any of the estimation techniques. Readers are directed to the fact that a single
computer code is not available for implementation of the various estimation
techniques described in this report.
In no case do the a priori techniques described in this report replace sound
engineering practice. A priori evaluation of alternatives is one thing. Selec-
tion, recommendation, and funding of a preferred alternative is quite another.
In the latter, the preferred alternative must stand on its own merit as environ-
mentally protective and cost-effective. Substantial sediment-specific testing is
usually required to clearly demonstrate that a given alternative is at once envi-
ronmentally protective and cost-effective. No amount of a priori estimation of
contaminant losses is sufficient for this task. This report provides a planning
level assessment tool for narrowing the universe of available alternatives and,
hence, the scope of sediment-specific testing required for decision making.
Nonremoval technologies
Estimating contaminant losses for nonremoval technologies is difficult due
to lack of field databases and standard procedures for assessment for non-
removal technologies. Predictive models based on diffusion are conceptually
applicable to most nonremoval technologies. However, predictive techniques
are not available that account for many important aspects of remediation with
nonremoval technologies.
Losses during placement of a cap, or injection of immobilization additives,
or injection of reagents for chemical treatment can result in highly
243
Chapter 11 Summary and Recommendations
-------
localized-transient disturbances of contaminated sediment. These highly local-
ized and transient disturbances can be as important, if not more important,
than long-term diffusion losses. At present, highly localized-transient losses
associated with nonremoval technology implementation cannot be predicted.
Once the implementation phase of a nonremoval technology is completed,
diffusion is the major loss pathway in the absence of significant advection.
Application of diffusion models to in situ capping is a recent development in
contaminant loss estimation. The theoretical basis for diffusion modeling is
well developed and confirmed in laboratory-scale simulations of in situ cap-
ping, but field verification data are nonexistent. Convection, bioturbation,
and biodegradation are potentially important, depending on site characteristics.
Convection and bioturbation effects can be avoided by careful planning,
design, and preproject testing. For example, controls for bioturbation should
be part of engineering design, and sites with significant groundwater move-
ment through the sediment are not good candidates for in situ capping.
Dredging
Techniques for estimating sediment solids losses during hydraulic and
mechanical dredging are available for conventional dredging equipment.
Techniques are not available for innovative dredging equipment options. The
available predictive techniques provide estimates of sediment losses in terms
of mass loss per time or mass loss per in situ volume dredged. Exposure
concentrations are not estimated. To estimate exposure concentrations, the
predicted losses of sediment and associated chemical contaminants must be
incorporated into water quality or exposure assessment models.
Techniques for estimating contaminant losses during dredging are still in an
early stage of development. Field data on turbidity and suspended solids
downstream of dredging operations are available, but measurement of losses at
the point of dredging that gave rise to the reported data are largely lacking.
Empirical correlations of sediment losses at the point of dredging with dredg-
ing operational parameters have been developed, but field validation data are
scarce. The predictive techniques focus on losses at the point of dredging and
are inherently a priori, although laboratory tests have been proposed. It is
anticipated that the available correlations will be modified and improved as a
result of ongoing studies.
Transportation
Techniques for estimating losses of sediments and associated chemical
contaminants during transportation of dredged material are not available for
most transportation modes. Pipeline breaks, scow spillage, and truck acci-
dents can be expected, but the frequency of such events have not been
documented, and there has been little effort to quantify the associated losses.
244
Chapter 11 Summary and Recommendations
-------
A priori predictive techniques for losses from scows due to volatilization are
available.
Pretreatment and disposal facilities
Key contaminant migration pathways, techniques for estimating losses, and
qualitative indications of predictive reliability for pretreatment and CDFs are
identified in Table 25. Pretreatment and confined disposal are remediation
components for which engineering controls on contaminant losses are most
practical.
Table 25
Availability and Reliability of Contaminant Loss Estimation Tech-
niques for Pretreatment and Confined Disposal Facilities
Pathway
Effluent
Hydraulic
Mechanical
Leachate
Volatilization
Runoff
Available
Yes
No
Yes
Yes
Yes
Reliability
High
Moderate
Low
High
Ease of Use
Simple
Simple but Complicated
Difficult
Simple
Contaminant migration pathways for pretreatment and CDFs are similar
because both facilities confine dredged material solids. There is always a
potential for leachate and volatile loss pathways to be of concern when consid-
ering pretreatment and confined disposal. In addition, hydraulic placement of
dredged material in pretreatment and CDFs will involve an effluent pathway.
The relative significance of these contaminant migration pathways is con-
taminant and facility design specific. Pathways involving movement of large
masses of water, such as effluent from hydraulic filling and long-term leach-
ing, have the greatest potential for moving significant quantities of soluble and
slightly soluble contaminants. Pathways such as volatilization may also result
in loss of organic chemicals during filling and storage.
A priori techniques are available for estimating losses via effluent, leach-
ate, and volatilization from pretreatment and CDFs. However, there are few
field verification data for the a priori techniques. For effluent resulting from
hydraulic filling, laboratory tests are available that have been field verified.
Confidence and accuracy in effluent predictions for hydraulic filling are conse-
quently high. There are no techniques for estimating losses during mechanical
filling of nearshore and in-water facilities.
Chapter 11 Summary and Recommendations
245
-------
Scientifically sound a priori and laboratory-based techniques are available
for estimating leachate quality. To estimate leachate losses, leachate quality
estimates must be coupled with computer models such as the Hydrologic Eval-
uation of Landfill Performance model to simulate site-specific hydrologic pro-
cesses (precipitation, evaporation, infiltration, percolation, etc.)- Leachate
prediction techniques have not been field verified. Confidence in predictions
is moderate relative to predictions for other contaminant migration pathways.
The only predictive techniques available for estimating volatile losses are a
priori techniques. In cases where highly contaminated dredged material is
disposed, volatile emissions should be evaluated to protect workers and others
who could inhale contaminants released through this pathway. The a priori
techniques were developed from chemical vapor equilibrium concepts and
transport phenomena fundamentals. Predicted emission rates are primarily
dependent on the chemical concentration in the dredged material, the surface
area through which emission occurs, and climatic factors such as wind speed.
Confidence in volatile emission calculations is low relative to predictions for
other contaminant migration pathways.
Dredged material treatment
Estimation of losses associated with dredged material treatment processing
follows standard engineering practice of conducting laboratory and pilot-scale
treatability studies. Performance data generated by treatability studies usually
provide the information on treatment process waste streams and residuals
needed to estimate losses and additional treatment requirements. The pilot-
scale treatability studies conducted in other elements of the ARCS program
may be used in planning level evaluations of treatment alternatives and associ-
ated contaminant losses. Caution should be exercised in using these data to
ensure that the treatment processes under consideration are applicable to the
sediment to be remediated. Sediment characteristics—physical and chemical—
and other site-specific factors can significantly affect implementability of a
treatment process.
Effluent/leachate treatment
Effluent and leachate may be viewed as wastewaters and as such are amen-
dable to conventional wastewater treatment processes. The available literature
on wastewater treatment engineering provides information suitable for plan-
ning level assessments of contaminant losses associated with effluent and
leachate treatment. As with dredged material treatment, before proceeding
with design and final engineering calculations including contaminant losses,
standard engineering practice involves conducting treatability studies. The
performance data generated by treatability studies usually provide the informa-
tion on treatment process waste streams and residuals needed to estimate
losses and additional treatment requirements.
246
Chapter 11 Summary and Recommendations
-------
Example calculations
Estimates of contaminant losses for components of a remediation alterna-
tive for a specific project can be pooled to provide an estimate for the entire
remediation alternative. Such estimates can then be used to rank alternatives
for remediation of a specific site. This approach has been illustrated through
a set of calculations for a sediment contaminated with PAHs in the Dead
Man's Creek portion of the Buffalo River, New York.
The example calculations show how to normalize losses with respect to the
volume of sediment to be remediated so that estimates for various pathways
can be pooled and alternatives can be compared on a common basis. The
remediation alternatives considered were as follows:
I Clamshell dredging with disposal in a CDF (with and without effluent
and leachate controls).
II Cutterhead dredging with disposal in a CDF (with and without effluent
and leachate controls).
Ill Clamshell dredging with stockpiling in a pretreatment facility
(with and without effluent and leachate controls) followed by thermal
desorption processing of the dredged material.
IV Cutterhead dredging with dewatering in a pretreatment facility (with
and without effluent and leachate controls) followed by thermal desorp-
tion processing of the dredged material.
V In Situ capping.
Example contaminant-loss calculations showed that in situ capping (Alter-
native V) were less than the losses for all other alternatives for remediation of
PAH-contaminated sediment in the Dead Man's Creek area of the Buffalo
River. PAH losses associated with in situ capping were estimated to be 1,000
to 100,000 times less than the next best alternative. Loss estimates for in situ
capping were significantly lower than any of the other alternatives because the
only contaminant migration pathway included in the analysis of in situ capping
was diffusion through the cap. Losses due to disturbance of contaminated
sediment during cap placement, release of excess pore pressure during consol-
idation, and erosion by extreme flow events were not estimated. Subject to
these limitations, the large difference between in situ capping and the other
alternatives suggests that in situ capping can be a very effective means of
sediment remediation. A cap can be armored to improve its stability, and in
situ capping would generally only be considered for sites subject to weak
erosive forces and no significant ground water movement.
Among Alternatives I through IV, Alternative IV with loss control mea-
sures, provided the least return of contaminants to the environment. There
was, however, very little difference in total contaminant losses between
247
Chapter 11 Summary and Recommendations
-------
disposal in a CDF and thermal desorption for low mobility contaminants such
as benzopyrene and benzanthracene. Thermal desorption rather than confined
disposal of benzanthracene, for example, provided only about 3 percent less
return of the contaminant to the environment.
Thermal desorption of the more mobile contaminant, phenanthrene, how-
ever, resulted in a reduction of total phenanthrene losses by 56 percent over
the CDF option. This comparison assumes that engineering controls over
effluent and leachate losses from pretreatment and CDFs are in place. In the
absence of such controls, the CDF option results in much higher contaminant
losses of phenanthrene. Absence of controls increased the losses during appli-
cation of the pretreatment/thermal desorption option by 56 percent, and the
losses for disposal in a CDF were more than an order of magnitude larger.
Differences were evident between low and high mobility contaminants
upon application of the clamshell or cutterhead dredging options. Clamshell
dredging tends to release a greater quantity of resuspended sediments com-
pared with hydraulic cutterhead dredging. Low-mobility contaminants such as
benzanthracene and benzopyrene are strongly associated with the sediment
particles and are therefore released in greater quantities by clamshell dredging.
High mobility contaminants such as phenanthrene and anthracene, however,
tend to exhibit greater losses during cutterhead dredging. The large quantities
of water needed to dredge hydraulically significantly increase the mass of
these compounds in the water phase and thus increase volatile, effluent and
leachate losses of these more soluble compounds. Comparison of Alterna-
tive I and II with loss control measures, for example, shows approximately
twice as much phenanthrene and anthracene lost during application of cutter-
head dredging than during application of clamshell dredging, while signifi-
cantly reducing losses for the less soluble benzanthracene and benzopyrene.
Although the magnitude of the losses during application of any of these
alternatives is sensitive to the particular set of assumptions employed, the
results clearly suggest that the optimum remedial alternative, that is the alter-
native leading to a minimum loss of contaminants, can be a strong function of
the particular contaminant. In addition, the presumption that greater control
of suspended paniculate losses leads to greater control of contaminant losses is
not entirely accurate.
Recommendations
Recommendations are provided for using contaminant loss estimates and
for research needed to improve the reliability and accuracy of available esti-
mation techniques and develop techniques where none are presently available.
94.R
*-^° Chapter 11 Summary and Recommendations
-------
Uses
The estimation techniques and the example approach to using contaminant
loss estimates described in this report were designed for comparison purposes.
Their best use is in relative comparison and ranking of alternatives and rela-
tive comparison of loss pathways for specific alternatives. Contaminant-loss
estimation exercises conducted solely to justify a predetermined preferred
alternative should be avoided. Specific recommendations for using
contaminant-loss estimates are provided below.
a. Contaminant-loss estimates should be used in a preproject planning
mode to help screen remediation alternatives. Contaminant-loss esti-
mates alone are not sufficient for decision making.
b. Contaminant-loss estimates for a specific alternative should be used to
indicate critical loss pathways where engineering controls should be
considered and potentially provide the most return.
c. Contaminant-loss estimates for a specific alternative should be used to
indicate where pathway specific laboratory testing is needed to improve
estimates and provide information for evaluating the feasibility of
engineering controls.
d. Contaminant-loss estimates should be used as input for risk assessment.
e. Contaminant-loss estimates for preferred alternatives should be used to
demonstrate the merit of the preferred alternative relative to the
no-action alternative and to indicate where engineering controls may
provide benefit relative to the no-action alternative.
Research needs
Research needs for improving available contaminant-loss estimation tech-
niques and providing estimation techniques where none are available are pro-
vided below in order of priority. The order of priority was developed within
the context of freshwater sediment remediation and not maintenance dredging
or remediation of estuarine sediments.
a. In situ capping—A priori and laboratory-based techniques for estimat-
ing contaminant losses during the implementation phase of an in situ
capping project are a top research priority. The available contaminant
loss estimation techniques for in situ capping account for diffusion
losses alone. Loss estimates for in situ capping will, therefore, gener-
ally be lower than estimates for most other alternatives. However,
implementation losses are probably much higher than diffusion losses
and, if accounted for, could dominant the loss estimates for in situ cap-
ping. Improved techniques that account for implementation losses are
needed to provide a more realistic preproject assessment of in situ
249
Chapter 11 Summary and Recommendations
-------
capping. Development of improved contaminant loss estimation tech-
niques for in situ capping should be conducted such that the techniques
are also applicable to capping dredged material. A field verification
component will also be needed.
b. Dredged Material Treatment—Treatability studies sometimes do not
provide sufficient data to fully evaluate contaminant losses.
Laboratory- and pilot-scale treatability studies in which detailed data
are obtained on all process streams are needed in order to fully evalu-
ate contaminant losses and develop a database for preproject evaluation
of probable contaminant losses. This information is needed to demon-
strate the relative merit of treatment to other alternatives. Because
treatability studies are very expensive, efforts should be made to obtain
as much process and contaminant loss information as possible. Treat-
ability studies that do not include a detailed materials balance should be
avoided.
c. Volatile Emissions—Volatile emissions are potentially important from a
worker health and safety viewpoint for highly contaminated sediments
and dredged materials, as well as, a potentially important contaminant
loss pathway. Research is needed to improve the available a priori
estimation techniques and develop laboratory-based estimation tech-
niques. A field verification component is also needed.
d. Leachate flow—The Hydrologic Evaluation of Landfill Performance
computer model is an excellent model for estimating leachate flow
from upland pretreatment and CDFs. Many of the assumptions on
which the model is based are not readily applicable to nearshore and
in-water facilities in which significant lateral seepage may occur. The
model can be "tricked" to simulate losses through dikes, but a model
designed to evaluate such problems would be preferred. A time-
varying contaminant transport model that simulates fluctuating water
levels and the attendant changes in hydraulic gradients is needed to
fully evaluate leachate seepage in nearshore and in-water facilities.
e. Effluent Losses—Available estimation techniques are limited to
hydraulic filling, and most of the laboratory and field data behind the
available techniques are for inorganic contaminants. Additional field
verification involving organic contaminants is needed to supplement the
data on inorganic contaminants and provide a complete picture of the
predictive capability of the modified elutriate test. In addition,
research and development are needed to provide laboratory and a priori
estimation techniques for effluent losses during mechanical filling of
nearshore and in-water pretreatment and disposal facilities.
250
Chapter 11 Summary and Recommendations
-------
References
Ambrose, R. B. et al. (1987). "WASP4, a hydrodynamic and water quality
model—Model theory, user's manual, and programmer's guide," EPA/600/
3-87-039, U.S. Environmental Protection Agency, Athens, GA.
Averett, D. E. (1988). "New Bedford Harbor Superfund Project, Acushnet
River estuary engineering feasibility study of dredging and dredged mate-
rial disposal alternatives; Report 3, Characterization and elutriate testing of
Acushnet River estuary sediment," Technical Report EL-88-15, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Averett, D. E., and Otis, M. J. (1990). "New Bedford Harbor Superfund
Project, Acushnet River estuary engineering feasibility study of dredging
and dredged material disposal alternatives; Report 12, Executive sum-
mary," Technical Report EL-88-15, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Averett, D. E., Palermo, M. R., Otis, M. J., and Rubinoff, P. B. (1988).
"New Bedford Harbor Superfund Project, Acushnet River estuary engi-
neering feasibility study of dredging and dredged material disposal alter-
natives; Report 11, Evaluation of conceptual dredging and disposal
alternatives," Technical Report EL-88-15, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Averett, D. E., Palermo, M. R., and Wade, R. (1988). "Verification of
procedures for designing dredged material containment areas for solids
retention," Technical Report D-88-2, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Averett, D. E., Perry, B. D., Torrey, E. J., and Miller, J. A. (1990).
"Review of removal, containment and treatment technologies for
remediation of contaminated sediment in the Great Lakes," Miscellaneous
Paper EL-90-25, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Bahr, J. M., and Rubin, J. (1987). "Direct comparison of kinetic and local
equilibrium formulations for solute transport affected by surface reactions,"
Water Resources Research 23(3), 438-452.
251
References
-------
Barnard, W. D. (1978). "Prediction and control of dredged material disper-
sion around dredging and open-water disposal operation," Technical Report
DS-78-13, U.S. Army Engineer Waterways Experiment Station, Vicks-
burg, MS.
Bear, J., and Verruijt, A. (1987). Modeling groundwater flow and pollution.
D. Reidel Publishing Company, Dordrecht, Holland.
Berger, B. B. (1987). Control of organic substances in water and
wastewater. Noyes Data Corporation, Park Ridge, NJ.
Bohlen, W. F. (1978). "Factors governing the distribution of dredge-
resuspended sediments," Proceedings, 16th Coastal Engineering Confer-
ence, ASCE 2001-2017.
Bohlen, W. F., and Tramontaro, J. M. (1977). "An investigation of the
impact of dredging operations on suspended material transport in the
Lower Thames River Estuary," University of Connecticut, Groton, CT.
Brandoma, M. G., and Divory, D. J. (1976). "Development of models for
prediction of short-term fate of dredged material discharged in the estuarine
environment," Contract Report D-76-5, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Brannon, J. M. (1978). "Evaluation of dredged material pollution potential,
synthesis report," Technical Report DS-78-6, U.S. Army Engineer Water-
ways Experiment Station, Vicksburg, MS.
Brannon, J. M., Engler, R. M., Rose, J. R., Hunt, P. G., and Smith, I.
(1976). "Selective analytical partitioning of sediments to evaluate potential
mobility of chemical constituents during dredging and disposal operation,"
Technical Report D-76-7, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Brannon, J. M., Hoeppel, R. E., Smith, I., and Gunnison, D. (1986).
"Long-term effectiveness of capping in isolating dutch kills sediment from
biota and the overlying water," Miscellaneous Paper EL-86-8, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Brannon, J. M., Hoeppel, R. E., Sturgis, T. C., Smith, I., and Gunnison, D.
(1985). "Effectiveness of capping in isolating contaminated dredged
material from biota and the overlying water," Technical Report D-85-10,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Brannon, J. M., Myers, T. E., Gunnison, D., and Price, C. B. (1991).
"Nonconstant polychlorinated biphenyl partitioning in New Bedford Harbor
Sediment during sequential batch leaching," Environmental Science and
Technology 25(6), 1082-1087.
9R9
f-'Jf- References
-------
Brannon, J. M., Myers, T. E., and Price, C. B. (1990). "Factors affecting
the behavior of metals in leachate," Environmental Effects of Dredging
Technical Notes, EEDP-02-13, U.S. Army Engineer Waterways Experi-
ment Station, Vicksburg, MS.
Brannon, J. M., Myers, T. E., and Price, C. B. (1992). "Leachate testing
of Hamlet City Lake North Carolina Sediment," Miscellaneous Paper
D-92-5, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Brannon, J. M., Myers, T. E., and Tardy, B. A. (1994). "Leachate testing
and evaluation for freshwater sediments," Miscellaneous Paper D-94-1,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Brannon, J. M., Pennington, J. C., Gunnison, D., and Myers, T. E. (1990).
"Comprehensive analysis of migration pathways (CAMP): Migration path-
ways controlling internal cycling and net mass transport of contaminants at
confined dredged material disposal facilities," Miscellaneous Paper D-90-5,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Brannon, J. M., Pennington, J. C., Myers, T. E., and Price, C. B. (1989).
"Factors affecting leachate quality," Environmental Effects of Dredging
Technical Notes, EEDP-02-11, U.S. Army Engineer Waterways Experi-
ment Station, Vicksburg, MS.
Brannon, J. M., Plumb, R. H., and Smith, I., Jr. (1980). "Long-term
release of heavy metals from sediments," Contaminants and sediments:
Volume II. R. A. Baker, ed., Ann Arbor Science, Ann Arbor, MI.
Burks, S. A., and Engler, R. M. (1978). "Water quality impacts of aquatic
dredged material disposal (laboratory investigations)," Technical Report
DS-78-4, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Burns, L. A., and Cline, D. M. (1985). Exposure analysis modeling system,
Reference Manual for EXAMS II. EPA-600/3-85-038, U.S. Environmental
Protection Agency, Athens, GA.
Carpenter, B. H. (1986). "PCB sediment decontamination-technical/
economic assessment of selected alternative treatments," Contract Report
68-02-3992, Research Triangle Institute, Research Triangle Park, NC.
Carslow, H. S., and Jaeger, J. C. (1959). Conduction of heat in solids.
Oxford University Press, Oxford, UK.
Chiou, C. T., Porter, P. E., and Schmedding, D. W. (1983). "Partition
equilibria of nonionic organic compounds between soil organic matter and
water," Environmental Science and Technology 17(4), 227-231.
253
References
-------
Cleary, R. W., and Adrian, D. D. (1973). "Analytical solution of the
convective-dispersive equation for cation adsorption in soils," Soil Science
Society of America Proceedings 37, 197-199.
Cohen, Y., Cocchio, W., and Mackay, D. (1978). "Laboratory study of
liquid-phase controlled volatilization rates in presence of wind waves,"
Environmental Science and Technology 12, 553-558.
Collins, M. A. (1989). "Dredging induced near field resuspended sediment
concentrations and source strengths," Draft Report on Contract
No. DAALO3-86-D-0001, Scientific Services Program for the U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Corbitt, R. A. (1989). Standard handbook of environmental engineering.
McGraw-Hill Inc., New York.
Crank, J. (1975). The mathematics of diffusion. Oxford University Press,
Oxford, UK.
Cullinane, M. J., Averett, D. E., Shafer, R. A., Male, J. W., Truitt, C. L.,
and Bradbury, M. R. (1986). "Guidelines for selecting control and treat-
ment options for contaminated dredged material requiring restrictions,"
Puget Sound Dredged Disposal Analysis Report, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
DiGiano, F. A., Miller, C. T., Yoon, J. (1993). "Predicting release of
PCBs at the point of dredging," Journal of Environmental Engineering,
American Society of Civil Engineers 119(1), 72-89.
. (1995). "Dredging elutriate test (DRET) development," Con-
tract Report D-95-1, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Dilling, W. L. (1977). "Interphase transfer processes. II. Evaporation rates
of chloromethanes, ethanes, ethylenes, propanes, and propylenes from
dilute aqueous solutions. Comparisons with theoretical predictions,"
Environmental Science and Technology 11, 405-409.
Dortch, M. S., Hales, L. Z., Letter, J. V., McAnally, W. H. (1990).
"Methods of determining the long-term fate of dredged material for aquatic
disposal sites," Technical Report D-90-1, U.S. Army Engineer Waterways
Experiment station, Vicksburg, MS.
Doyle, R. D., Myers, T. E., and Adrian, D. D. (1993). "Atrazine retention
and disappearance in experimental wetlands mesocosms," Presented at
Wetlands Biogeochemical Symposium, Louisiana State University,
February 1993.
References
-------
Dupont, R. R. (1986). "Evaluation of air emission release rate model
predictions of hazardous organics from land treatment facilities," Environ-
mental Progress 5(3), 197.
EBASCO Services Incorporated. (1990). "Final air monitoring report
Volume I, New Bedford Harbor Pilot Dredging and Disposal Study," Con-
tractor Report for EPA Contract Number 68-01-7250.
Engler, R. M., Patin, T. R., and Theriot, R. F. (1990). "Update of the
Corps' environmental effects of dredging programs (FY 89)," Miscella-
neous Paper D-90-2, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Environmental Laboratory. (1987). "Disposal alternatives for PCB-
contaminated sediments from Indiana Harbor, Indiana," Miscellaneous
Paper EL-87-9, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Environmental Laboratory, Department of the Army, Waterway Experiment
Station, Corps of Engineers. (1989). CE-QUAL-RIVI: A Dynamic, One-
dimensional (Longitudinal), Water Quality Model for Streams: User's
Manual.
Francingues, N. R., Jr., Palermo, M. R., Lee, C. R., and Peddicord, R. K.
(1985). "Management strategy for disposal of dredged material: Contami-
nant testing and controls," Miscellaneous Paper D-85-1, U.S. Army Engi-
neer Waterways Experiment Station, Vicksburg, MS.
Fred, D. L. (1988). The NWS DAMBRK Model Theoretical Background/
User Documentation. Hydrologic Research Laboratory, Office of Hydrol-
ogy, National Weather Service.
Freeze, R. A., and Cherry, J. A. (1979). Groundwater. Prentice-Hall, Inc.
Englewood Cliffs, NJ.
Gambrell, R. P., Khalid, R. A., and Patrick, W. H. (1978). "Disposal
alternatives for contaminated dredged material as a management tool to
minimize adverse environmental effects," Technical Report DS-78-8,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Glover, R. E. (1974). "Transient ground water hydraulics," Water
Resources Publications, Littleton, CO.
Halliburton, T. A. (1978). "Guidelines for dewatering/densifying confined
dredged material," Technical Report DS-78-11, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
References
-------
Hamaker, J. W. (1972). "Diffusion and volatilization," Organic chemicals
in the soil environment, Volume I. E. A. I. Goring and J. W. Hamaker,
ed., Marcel Dekker, New York.
Hammer, D. A., ed. (1989). Constructed wetlands for wastewater treatment.
Lewis Publishers, Chelsea, MI.
Harr, M. E. (1962). Groundwater and seepage. McGraw-Hill Book
Company, New York.
Hayes, D. F. (1986). "Development of a near field source strength model to
predict sediment resuspension from cutter suction dredges," M.S. thesis,
Department of Civil Engineering, Mississippi State University, MS.
Hayes, D. F., McLellan, T. N., and Truitt, C. L. (1988). "Demonstrations
of innovative and conventional dredging equipment at Calumet Harbor, Illi-
nois," Miscellaneous Paper EL-88-1, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Hayes, D. F., Raymond, G., McLellan, T. N. (1984). "Sediment resuspen-
sion from dredging activities." Proceedings, ASCE Specialty Conference,
Dredging '84. American Society of Civil Engineers, Clearwater Beach,
FL.
Hayter, E. J. (1987). "Finite element hydrodynamic and cohesive sediment
transport modeling system," Department of Civil Engineering, Clemson
University.
Herbich, J. B., and Brahme, S. B. (1991). "Literature review and technical
evaluation of sediment resuspension during dredging," Contract Report
HL-91-1, prepared by Texas A&M University, College Station, TX, for
the U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Hill, D. O., Myers, T. E., and Brannon, J. M. (1988). "Development and
application of techniques for predicting leachate quality in confined dispo-
sal facilities: Background and theory," Miscellaneous Paper D-88-1,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Hoeppel, R. E., Myers, T. E., and Engler, R. M. (1978). "Physical and
chemical characterization of dredged material influents and effluents in
confined land disposal areas," Technical Report D-78-24, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Hosseinipour, E. Z., and Martin, J. L. RIVMOD: (1991). A one-
dimensional hydrodynamic and sediment transport model—Model theory
and user's manual, AScI Corporation at U.S. Environmental Protection
Agency, Athens, GA.
256
References
-------
Huston, W., and Huston, W. C. (1976). "Techniques for reducing turbidity
with present dredging procedures and operations," Contract Report D-76-
4, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Ikalainen, A. J. (1987). "Detailed analysis of remedial technologies for the
New Bedford Harbor feasibility study," Contract Report 68-01-7250,
Ebasco Services, Inc.
Jennings, A. A., and Kirkner, J. J. (1984). "Instantaneous equilibrium
approximation analysis," Journal of Hydraulic Engineering 110(12), 1700-
1717.
Johnson, B. H. (1990). "User's guide for models of dredged material dispo-
sal in open water," Technical Report D-90-5, U.S. Army Engineer Water-
ways Experiment Station, Vicksburg, MS.
Jones, R. A., and Lee, G. F. (1978). "Evaluation of the elutriate test as a
method of predicting contaminant release during open-water disposal of
dredged sediments and environmental impact of open-water dredged mate-
rial disposal Vol. I: Discussion," Technical Report D-78-45, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Khan, A., and Grossi, J. (1984). "Confined disposal in the Canadian Great
Lakes," Dredging and dredged material disposal. R. L. Montgomery and
J. W. Leach, ed., American Society of Civil Engineers, New York, 654-
663.
Karickhoff, S. W. (1981). "Semi-empirical estimation of sorption of hydro-
phobic pollutants on natural sediments and soils," Chemosphere 10(9),
833-846.
Karickhoff, S. W., Brown, D. S., and Scott, T. A. (1979). "Sorption of
hydrophobic pollutants on natural sediments," Water Research 13, 241-
248.
Kiang, U., and Metry, A. R. (1982). Hazardous waste processing technol-
ogy. Ann Arbor Science Publishers, Inc., Ann Arbor, MI.
Krizek, R. J., FitzPatrick, J. A., and Atmatzidia, D. K. (1976). "Investiga-
tion of effluent filtering systems for dredged material containment
facilities," Contract Report D-76-8, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Krizek, R. J., Gallagher, G. J., Karadi, G. N. (1976). "Water quality
effects of a dredging disposal area," Journal of the Environmental
Engineering Division, American Society of Civil Engineers 102(EE2), 389-
409.
257
References
-------
Kuppusamy, T. (1991). "Multiphase flow and transport through soils,"
Short Course, Nov 25-27, 1991, Vicksburg, MS.
Lee, C. R., and Skogerboe, J. G. (1984). "Restoration of pyritic soils and
the quantification of erosion control," Miscellaneous Paper EL-84-7,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Lee, C. R., Tatem, H. E., Brandon, D. L., Kay, S. H., Peddicord, R. K.,
Palermo, M. R., and Francingues, N. R., Jr. (1991). "General decision-
making framework for management of dredged materials," Miscellaneous
Paper D-91-1, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Leonard, R. P. (1988). "Surface and groundwater monitoring at Times
Beach confined disposal facility, Buffalo, New York." Proceedings of a
Seminar on Water Quality R&D: Successful Bridging Between Theory and
Applications. R. G. Willey, ed., U.S. Army Corps of Engineers Com-
mittee on Water Quality, Washington, DC.
Liss, P. S., and Slater, P. G. (1974). "Flux of gases across the air-water
interface," Nature 247, 181-184.
Louisiana Water Resources Research Institute. (1990). "Synthesis of
research needs workshop: Development of leach tests for contaminated
dredged material," Miscellaneous Paper D-90-3, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Lu, J. C. S., Eichenberger, B., Knezevic, M., and Chen, K. Y. (1978).
"Characterization of confined disposal area influent and effluent paniculate
and petroleum fractions," Technical Report D-78-16, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Lunney, P. D., Springer, C., and Thibodeaux, L. J. (1985). "Liquid-phase
mass transfer coefficients for surface impoundments," Environmental Prog-
ress 109(4), 203-211.
Lyman, W. J., Reehl, W. F., and Rosenblatt, D. H. (1990). Handbook of
chemical property estimation methods. American Chemical Society,
Washington, DC.
Mackay, D. (1979). "Finding fugacity feasible," Environmental Science and
Technology 13(10), 1218-1223.
MacKnight, S., and MacLellan, E. (1984). "PCB contaminated sediments
from Petit-de-Grat, Canada." Dredging and dredged material disposal.
R. L. Montgomery and J. W. Leach, ed., American Society of Civil
Engineers, New York, 619-624.
References
-------
Mang, J. L., Lu, J. C. S., Lofy, R. J., and Stearns, P. (1978). "A study of
leachate from dredged material in upland areas and/or in productive uses,"
Technical Report D-78-20, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Martin, J. L., Ambrose, R., and McCutcheon, S. (1988). "Assessment
protocols for confined disposal facilities: Time Variable TOXI4 Based
CDF Model," USEPA Environmental Research Laboratory, Athens, GA.
Mayer, R., Letey, J., and Farmer, W. J. (1974). "Models for predicting
volatilization of soil-incorporated pesticides," Soil Science Society of
America Proceeding 38, 563-568.
Maynord, S. (1993). "Design considerations for capping/armoring of
contaminated sediments in-place," Draft Technical Note prepared for Engi-
neering Technology Work Group, Assessment and Remediation of Contam-
inated Sediments Program, Great Lakes National Program Office,
U.S. Environmental Protection Agency, Chicago, IL.
McLellan, T. N., Havis, R. N., Hayes D. F., and Raymond, G. L. (1989).
"Field studies of sediment resuspension characteristics of selected
dredges," Technical Report HL-89-9, U.S. Army Waterways Experiment
Station, Vicksburg, MS.
McWhorter, D. B., and Sunada, D. K. (1977). Groundwater hydrology and
hydraulics. Water Resources Publications, Littleton, CO.
Means, J. C., Wood, S. G., Hassett, J. J., and Banwart, W. L. (1980).
"Sorption of polynuclear aromatic hydrocarbons by sediments and soils,"
Environmental Science and Technology 14(12), 1524-1528.
Millington, R. J., and Quirk, J. P. (1961). Transactions of the Faraday
Society 57, 1200-1207.
Montgomery, R. L. (1978). "Methodology for design of fine-grained
dredged material containment areas for solids retention," Technical Report
D-76-56, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Montgomery, R. L., Thackston, E. L., and Parker, F. L. (1983). "Dredged
material sedimentation basin design," Journal of the Environmental Engi-
neering Division, American Society of Civil Engineers 109(2).
Morin, J., Cluff, D. B., and Powers, W. B. (1970). "Realistic rainfall simu-
lation for field investigation," presented at the 51st Meeting of the Ameri-
can Geophysical Union, Washington, DC.
259
References
-------
Morin, J., Goldberg, D., and Seginer, I. (1967). "A rainfall simulator
with a rotating disk," Transactions of the American Society of Agricul-
tural Engineers 10(1), 74-79.
Myers, T. E. (1991). "Polychlorinated biphenyl levels in the Saginaw con-
fined disposal sacility during disposal operations, Fall 1987," Miscel-
laneous Paper EL-91-4, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Myers, T. E., and Brannon, J. M. (1988). "Evaluation of leachate quality,"
Report 5 of 12, New Bedford Harbor Superfund Project: Acushnet River
Estuary Engineering Feasibility Study Series, Technical Report E-88-15,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Myers, T. E., Brannon, J. M., and Price, C. B. (1992). "Recent develop-
ments in leachate testing and evaluation," Miscellaneous Paper D-92-2,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Myers, T. E., Gambrell, R. P., and Tittlebaum, M. E. (1991). "Design of
an improved column leaching apparatus for sediments and dredged
material," Miscellaneous Paper D-91-3, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Myers, T. E., Stark, T. D., Gibson, A. C., Dardeau, E. A., Jr., and
Schroeder, P. R. (1993). "Management plan for the disposal of
contaminated dredged material in the Craney Island dredged material man-
agement area," Technical Report EL-93-20, U.S. Army Engineer Water-
ways Experiment Station, Vicksburg, MS.
National Research Council. (1990). Ground water models: Scientific and
regulatory applications. National Academy Press, Washington, DC.
Nawrocki, M. A. (1974). "Demonstration of the separation and disposal of
concentrated sediment," EPA 600/2-74-072, U.S. EPA, Washington, DC.
O'Connor, J. M., and O'Conner, S. G. (1983). "Evaluation of the 1980
capping operations at the experimental mud dump site, New York Bight
Apex," Technical Report D-83-3, U.S. Army Engineer Waterways Experi-
ment Station, Vicksburg, MS.
Ofuji, I., and Naoshi, I. (1976). "Antiturbidity overflow system for hopper
dredge," Dredging, Environmental Effects and Technology: Proceedings of
the Seventh World Dredging Conference, WODCON VII, San Francisco,
CA. 207-234.
Palermo, M. R. (1986). "Development of a modified elutriate test for esti-
mating the quality of effluent from dredged material disposal areas," Tech-
nical Report D-86-4, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
References
-------
Palermo, M. R. (1988). "Field evaluations of the quality of effluent from
confined dredged material disposal areas," Technical Report D-88-1,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
. (1991a). "Design requirements for capping," Technical Note
DRP-5-03, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
. (1991b). "Site selection considerations for capping," Technical
Note DRP-5-04, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
. (1991c). "Equipment and placement techniques for capping,"
Technical Note DRP-5-05, U.S. Army Engineer Waterways Experiment
Station, Vicksburg, MS.
Palermo, M. R., Fredette, T., and Randall, R. E. (1992). "Monitoring con-
siderations for capping," Technical Note DRP-5-07, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Palermo, M. R., Homziak, J., and Teeter, A. M. (1990). "Evaluation of
clamshell dredging and barge overflow, Military Ocean Terminal Sunny
Point, North Carolina," Technical Report D-90-6, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Palermo, M. R., Maynard, S., Miller, J., and Reible, D. D. "Guidance for
in situ subaqueous capping of contaminated sediments," Report in prepara-
tion, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Palermo, M. R., Montgomery, R. L., and Poindexter, M. E. (1978).
"Guidelines for designing, operating, and managing dredged material
containment areas," Technical Report, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Palermo, M. R., and Pankow, V. R. (1988). "New Bedford Harbor Super-
fund Project, Acushnet River Estuary Engineering Feasibility Study of
Dredging and Dredged Material Disposal Alternatives; Report 10,
Evaluation of Dredging and Dredging Control Technologies," Technical
Report EL-88-15, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Palermo, M. R., Shafer, R. A., Brannon, J. M., Myers, T. E., Truitt, C. L.,
Zappi, M. E., Skogeboe, J. G., Sturgis, T. C., Wade, R., Gunnison, D.,
Griffin, D. M., Jr., Tatum, H., Portzer, S., and Adamec, S. A. (1989).
"Evaluation of dredged material disposal alternatives for U.S. Navy home-
port at Everett, Washington," Technical Report EL-89-1, U.S. Army Engi-
neer Waterways Experiment Station, Vicksburg, MS.
261
References
-------
Palermo, M. R., and Thackston, E. L. (1988a). "Verification of predictions
of dredged material effluent quality," Journal of Environmental Division,
American Society of Civil Engineers 114(6).
Palermo, M. R., and Thackston, E. L. (1988b). "Test for dredged material
effluent quality," Journal of Environmental Division, American Society of
Civil Engineers 114(6).
. (1988c). "Flocculent settling above the Zone Settling Inter-
face," Journal of Environmental Division, American Society of Civil Engi-
neering 114(4).
. (1989). "Refinement of column settling test procedures for
estimating the quality of effluent from confined disposal areas," Technical
Report D-89-9, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Palermo, M. R., Zappi, P. A., Dillon, T. M., McFarland, V. A., Reilly,
F. J., Jr., Moore, D. D., Myers, T. E., Schneffner, N. W., Hales, L. Z.,
and Thackston, E. L. (1993). "Long-term management strategy for
dredged material disposal for the Naval Weapons Station, Yorktown, VA;
Naval Supply Center, Cheatham Annex, Williamsburg, VA; and Naval
Amphibious Base, Little Creek, Norfolk, VA," Miscellaneous Paper
EL-93-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Peddicord, R. K. (1988). "Summary of the U.S. Army Corps of
Engineers/U.S. Environmental Protection Agency Field Verification
Program," Technical Report D-88-6, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
Peters, R. E., Lee, C. R., and Bates, D. J. (1981). "Field investigations of
overland flow treatment of municipal lagoon effluent," Technical Report
EL-81-9, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Phillips, R. C., Westerdahl, H. E., Mize, A. L., and Robinson, S. A.
(1993). "Summary of literature describing the functional ability of wet-
lands to enhance wastewater quality," Technical Report WRP-CP-2,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Poindexter-Rollings, M. E. (1990). "Methodology for analysis of
subaqueous sediment mounds," Technical Report D-90-2, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Ponnamperuma, F. N. (1972). "The chemistry of submerged soils,"
Advances In Agronomy 24, 29-96.
References
-------
Reed, S. C. (1990). Natural systems for wastewater treatment: Manual of
practice. Water Pollution Control Federation, Alexandria, VA.
Reid, R. C., Prausnitz, J. M., and Sherwood, J. K. (1977). The properties
of gases and liquids. McGraw-Hill Book Company, New York.
Richardson, T. W., Kite, J. E., Jr., Schafer, R. A., and Ethridge, J. D., Jr.
(1982). "Pumping performance and turbidity generation of model 600/100
pneuma pump," Technical Report HL-82-8, U.S. Army Engineer Water-
ways Experiment Station, Vicksburg, MS. '
Rogers, J. H., Jr., and Dunn, A. (1992). "Developing design guidelines for
constructed wetlands to remove pesticides from agricultural runoff," Eco-
logical Engineering 1(1/2), 83-96.
Sato, E. (1976a). "Application of dredging techniques for environmental
problems," Proceedings ofWODCON VII: Dredging Environmental Effects
and Technology, 143-162.
Sato, E. (1976b). "Dredging techniques applied to environmental problems,"
World Dredging and Marine Construction 12(12), 32-38.
Schroeder, P. R. (1983). "Chemical clarification methods for confined
dredged material disposal," Technical Report D-83-2, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Schroeder, P. R., and Palermo, M. R. (1990). "The automated dredging and
disposal alternatives management system (ADDAMS)," Environmental
Effects of Dredging Technical Notes, EEDP-06-12, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Schroeder, P. R., and Peyton, R. L. (1987). "Verification of the hydraulic
evaluation of landfill performance (HELP) model using field data," EPA
600/2-87-050, U.S. Environmental Protection Agency, Cincinnati, OH.
Schroeder, P. R., Peyton, R. L., McEnroe, B. M., and Sjostrom, J. W.
(1988). "The hydrologic evaluation of landfill performance (HELP) model
Volume III. User's guide for Version 2," Unpublished manuscript,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Schwarzenbach, R. P., and Westall, J. (1981). "Transport of nonpolar
organic compounds from surface water to groundwater; Laboratory Sorp-
tion Studies," Environmental Science and Technology 15(11), 1360-1367.
Semmler, J. A. (1990). "PCB-volatilization from dredged material, Indiana
Harbor, IN," Environmental Effects of Dredging Technical Notes, EEDP-
02-12, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
263
References
-------
Semmler, J. A. (1993). "Application of equilibrium partitioning theory to a
confined disposal facility," M.S. thesis, Illinois Institute of Technology,
Chicago, IL.
Sheng, Y. P., Eliason, D. E. (1991). "A three-dimensional numerical model
of hydrodynamic and sediment transport in lakes and estuaries," Coastal
and Oceanography Engineering Department, University of Florida, for
U.S. Environmental Protection Agency, Athens, GA.
Shields, F. D., Jr., Thackston, E. L., Schroeder, P. R., and Bach, D. P.
(1987). "Design and management of dredged material containment areas
to improved hydraulic performance," Technical Report D-87-2, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Shuckrow, A. J., Pajak, A. P., and Osheka, J. W. (1981). "Concentration
technologies for hazardous aqueous waste treatment," EPA-600/2-81-019.
Skogerboe, J. G., Lee, C. R., Price, R. A., Brandon, D., and Hollins, G.
(1987). "Prediction of surface runoff water quality from Black Rock
Harbor dredged material placed in an upland site," Miscellaneous Paper
D-87-1, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Skogerboe, J. G., Price, R. A., and Brandon, D. L. (1988). "New Bedford
Harbor Superfund Project, Acushnet River Estuary Engineering Feasibility
Study of Dredging and Dredged Material Disposal Alternatives; Report 4,
Surface Runoff Quality Evaluation For Confined Disposal," Technical
Report EL-88-15, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Stenneker, R. C. H., van der Sloot, H. A., and Das, H. A. (1988). "Leach-
ing studies on dredged material in oxidized and reduced state," The Science
of the Total Environment 68, 11-23.
Sturgis, T. C., and Gunnison, D. (1988). "New Bedford Harbor Superfund
Project, Acushnet River Estuary Engineering Feasibility Study of Dredging
and Dredged Material Disposal Alternatives; Report 6, Laboratory Testing
for Subaqueous Caping," Technical Report EL-88-15, U.S. Army Engineer
Waterways Experiment Station, Vicksburg, MS.
Suarez, L. A. et al. (1986). GETS, a Simulation Model for Dynamic Bioac-
cumulation of Nonpolar Organics by Gill Exchange: A User's Guide.
U.S. Environmental Protection Agency, Athens, GA, EPA/600/S3-86/057.
Taylor, A. W., and Glotfelty, D. E. (1988). "Evaporation from soils and
crops," Environmental chemistry of herbicides Volume I. R. Grover, ed.,
CRC Press, Boca Raton, FL.
264
References
-------
Thackston, E. L., and Palermo, M. R. (1990). "Field evaluation of the
quality of effluent confined disposal areas supplemental study—Houston
Ship Channel," Technical Report D-90-9, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, MS.
. (1992). "Predicting effluent PCBs from superfund site dredged
material," Journal of Environmental Engineering, American Society of
Civil Engineers 118(5), 657-665.
Thibodeaux, L. J. (1979). Chemodynamics. Wiley Interscience, New York.
. (1989). "Theoretical models for evaluation of volatile emis-
sions to air during dredged material disposal with applications to
New Bedford Harbor, MA," Miscellaneous Paper EL-89-3, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
. (1990). "A theoretical evaluation of the effectiveness of cap-
ping PCB contaminated New Bedford Harbor bed sediment," Report pre-
pared for Balsam Environmental Consultants, Inc., Salem, NH.
Thibodeaux, L. J., and Hwang, S. T. (1982). "Landfarming of petroleum
waste modeling the air pollution problem," Environmental Progress 1(1),
42.
Thibodeaux, L. J., Parker, D. G., and Heck, H. H. (1984). "Chemical
emissions from surface impoundments," Environmental Progress 3(2), 73-
78.
Thoma, G. J., Reible, D. D., Valsaraj, K. T., and Thibodeaux, L. J. (1993).
"Efficiency of capping contaminated sediments in situ. 2. Mathematics of
diffusion-adsorption in the capping layer," Environmental Science and
Technology 27(12), 2412-2419.
Thomas, R. G. (1990a). "Volatilization from water," Handbook of chemical
property estimation methods. W. J. Lyman, W. F. Reehl, and D. H.
Rosenblatt, ed., American Chemical Society, Washington, DC.
. (1990b). "Volatilization from soil," Handbook of chemical
property estimation methods. W. J. Lyman, W. F. Reehl, and D. H.
Rosenblatt, ed., American Chemical Society, Washington, DC.
Truitt, C. (1987a). "Engineering considerations for subaqueous dredged
material capping—background and preliminary planning," Environmental
Effects of Dredging Programs Technical Notes, EEDP-01-3, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
OC C
References
-------
Truitt, C. (1987b). "Engineering considerations for capping subaqueous
dredged material deposits—design concepts and placement techniques,"
Environmental Effects of Dredging Programs Technical Notes,
EEDP-01-4, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Truitt, C. L. (1986). "The Duwamish Waterway Capping Demonstration
Project: Engineering analysis and results of physical monitoring," Techni-
cal Report D-86-2, U.S. Army Engineer Waterways Experiment Station,
Vicksburg, MS.
Turner, F. T., and Patrick, W. H., Jr. (1968). "Chemical changes in water
logged soils as a result of oxygen depletion," Proc. 9th International Con-
gress of Soil Science, Adelaide, Australia.
U.S. Army Corps of Engineers (USAGE). (1987). "Confined disposal of
dredged material engineer manual," EM 1110-2-5027, U.S. Army Corps
of Engineers, Washington, DC.
U.S. Army Engineer District, Buffalo. (1969). "Dredging and water quality
problems in the Great Lakes (in 12 volumes)," Buffalo, NY.
. (1993). Pilot-scale demonstration thermal desorption for the
treatment of Buffalo River sediments. EPA905-R93-005. Prepared for
U.S. Environmental Protection Agency, Great Lakes National Program
Office, Chicago, IL.
U.S. Environmental Protection Agency (USEPA). (1988). "Five year pro-
gram strategy for the Great Lakes National Program Office FY 1989-
1993," EPA Document No. 905/9/89/001, GLNPO Report No. 1-89,
Great Lakes National Program Office, U.S. Environmental Protection
Agency, Chicago, IL.
USEPA. (1989). "Guide for conducting treatability studies under CERCLA,
interim final," EPA/540/2-89/058, Office of Emergency and Remedial
Response, Washington, DC, and Office of Research and Development,
Cincinnati, OH.
(1992). Treatability Database, Version 4.0, Risk Reduction
Engineering Laboratory, Cincinnati, OH.
. (1993a). ARCS Risk Assessment and Modeling Overview Doc-
ument, EPA905-R93-007, USEPA Great Lakes National Program Office,
Chicago, IL.
. (1993b). "The superfund innovative technology evaluation pro-
gram: Technology profiles sixth edition," Superfund EPA/540/r-93/326,
November 1993, Office of Research and Development, U.S. Environmen-
tal Protection Agency, Washington, DC.
266 References
-------
USEPA. (1993c). "VISITT-vendor information system for innovative tech-
nologies summary report," EPA/540/0-00/000, Office of Research and
Development, U.S. Environmental Protection Agency, Washington, DC.
. (1994a). ARCS Remediation Guidance Document, EPA905-
B94-003, USEPA Great Lakes National Program Office, Chicago, IL.
. (1994b). ARCS Assessment Guidance Document, EPA905-
B94-002, USEPA Great Lakes National Program Office, Chicago, IL.
. (1994c). North American Wetlands for Water Quality Treat-
ment Database, Version 1.0, Risk Reduction Engineering Laboratory,
U.S. Environmental Protection Agency, Cincinnati, OH and Environmental
Research Laboratory, U.S. Environmental Protection Agency, Duluth,
MN.
USEPA/USACE. (1992). "Evaluating environmental effects of dredged
material management alternatives—A technical framework," EPA 842-B-
92-008, U.S. Environmental Protection Agency, Washington, DC.
Valocchi, A. J. (1985). Validity of the local equilibrium assumption for
modeling sorbing solute transport through homogenous soils. Water
Resources Research 21(6), 808-820.
Wang, X. Q., Thibodeaux, L. J., Valsaraj, K. T., and Reible, D. D. (1991).
"The efficiency of capping contaminated bed sediments in situ 1. Lab-scale
experiments on diffusion/adsorption in the capping layer," Environmental
Science and Technology 25(9), 1578-1584.
Wastewater Technology Centre. (1993). "Sediment treatment technologies
database, 2nd edition," Site Remediation Division, Wastewater Technology
Centre, Burlington, Ontario.
Westerdahl, H. E., and Skogerboe, J. G. (1982). "Realistic rainfall and
watershed response simulations for assessing water quality impacts of land
use management." Modeling Components ofHydrologic Cycle, Inter-
national Symposium on Rainfall-Runoff Modeling, Mississippi State Uni-
versity, Mississippi State, MS, 87-104.
Wischmeier, W. H., Johnson, C. B., and Cross, B. V. (1971). "A soil
erodibility nomograph for farmland and construction sites," Journal of Soil
and Water Conservation 21(5), 189-193.
Yagi, T. et al. (1975). "Effect of operating conditions of hydraulic dredges
on dredging capacity and turbidity," Translation from Japanese, No. 228,
Port and Harbor Research Institute, Ministry of Transport, Yokusaka,
Japan.
References
-------
Yagi, T. et al. (1977). "Turbidity caused by dredging." Dredging, Environ-
mental Effects and Technology: Proceeding of the Seventh World Dredging
Conference, WODCON VII, San Francisco, CA. 1079-1109.
Yu, K. Y., Chen, K. Y., Morrison, R. D., and Mang, J. L. (1978). "Physi-
cal and chemical characterization of dredged material sediments and
leachates in confined land disposal areas," Technical Report D-78-20,
U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Zappi, P. A., and Hayes, D. F. (1991). "Innovative technologies for dredg-
ing contaminated sediments," Miscellaneous Paper EL-91-20, U.S. Army
Engineer Waterways Experimental Station, Vicksburg, MS.
References
-------
Appendix A
Notation
a empirical swing velocity significance factor
A dredging area or area available for mass transfer
Ac capped area
Av surface area of vessel
Aw water surface area
b empirical tangential velocity significance factor
B Collins bucket parameter
C solubility in water
CEF contaminant containment efficiency factor for the
effluent pathway
CEFEFF containment efficiency based on effluent pathway only
Ca dissolved concentration of chemical in air
Cai background concentration of chemical in air at the dredged
material-air interface
CA water concentration of A
CAO pore water concentration in original sediment
Cc colloid concentration in water (DOC)
Q dissolved chemical concentration in water
C'd hypothetical dissolved chemical concentration in equilibrium with
background air
Cdoc colloidal specie concentration
CEFF.TOT total concentration of contaminant i in effluent
CINF.TOT tota^ concentration of contaminant i in influent
Cp suspended solids concentration
Cps suspended solids contaminant concentration
Cpw pore water concentration of A
Cpw° Pore water concentration in original sediment
Cs contaminant concentration in the sediment
Csorb concentration of contaminant sorbed to solid phase
CsL leachable metal concentration in dredged solids
C total whole water contaminant concentration
Cw aqueous phase contaminant concentration
C" background water concentration above cap
Cw* hypothetical dissolved chemical concentration in equilibrium with
background air
A i
Appendix A Notation
-------
C*w water concentration of A
d effective diameter of sediment grains
dp particle diameter
D depth of dredging
D' ratio of colloidal specie diffusivity to
DA molecular diffusivity (1 = air, 2 = water, 3 = sediment pore
spaces)
DA1 molecular diffusivity of chemical A in air
DA2 diffusivity of A in water
DA3 effective diffusivity of chemical A
DB molecular diffusivity of chemical B in water
Db effective biotubation diffusion coefficient
Dch diameter of cutterhead
D^ effective diffusivity, bracketed term equation
DF fractional depth of cut as a function of cutterhead diameter
Dp dispersion coefficient
Dv effective diameter of the vessel
EA3 effective dispersion coefficient in the medium
fb fraction of bucket dredge cycle on bottom
fd fraction of bucket dredge cycle for bucket insertion
f0 fraction of bucket dredge cycle out of water
fr fraction of sediment resuspended during dredging
fu fraction of bucket dredge cycle for bucket withdrawal
FD cutterhead resuspension rate factor accounting for degree of burial
FF cutterhead resuspension rate factor accounting for other factors
foc fraction organic carbon in sediment or dredged material
hb water depth (bucket dredging)
hj pond water elevation above base of dike
h2 outside water elevation above base of dike
H Henry's constant
Hj head at crown of water table mound
H2 head outside the confined disposal facility (CDF)
Hch height of cutterhead
i contaminant index
K hydraulic conductivity of the dike
KA1 overall mass transfer coefficient based on air-side concentrations
KA2 overall mass transfer coefficient based on water-side
concentrations
Kb benthic mass transfer coefficient
Kc colloid-water partition coefficient of A
Kd contaminant-specific equilibrium distribution
coefficient
KG gas-side mass transfer coefficient
KL liquid-side mass transfer coefficient
Ko colloid-water partition coefficient
Koc colloid-water partition coefficient of A
KOG overall gas-side mass transfer coefficient
KOL overall liquid-phase mass transfer coefficient
A2
Appendix A Notation
-------
Kov overall mass transfer coefficient
L horizontal distance separating surface of pond and surface of out-
side water body
Lbc characteristic length of clamshell bucket
LBio bioturbation layer thickness
Lcap effective depth of cap diffusive layer
Lch length of cutterhead
Lv vessel length
m mass of contaminant released
M weight of activated carbon
MA molecular weight of chemical A
MB molecular weight of chemical B
Mcs mass of contaminant in the solid phase
Mm mass of contaminant in the aqueous phase
Ms mass of solids
Mw mass of water
ne instantaneous flux of chemical A through the dredged material
nEDM instantaneous flux of chemical A through the dredged material-air
interface at time t
NA flux of contaminant A in free water
Nss steady-state flux
A^ flux through air-water interface
pA background partial pressure of chemical A in air
pA* partial pressure of component A in air at exposed surface (in
equilibrium with dredged material)
p*A pure component vapor pressure of chemical A
pA1 partial pressure of component A in background air
P total atmospheric pressure
q discharge per unit length of dike
Q volumetric flow of water
Qd volumetric flow of water through the averaging volume
R universal gas constant, 82.1 atm
Rj distance from center of CDF to edge of water table crown
R2 distance from center of CDF to the dike
RA release rate of contaminant
RA(t) release rate of contaminant, at time t
RA(t-*oo) release rate of contaminant, at steady-state
RD rate of contaminant release
RDb contaminant release rate
RD ch dissolved contaminant release rate for a cutterhead dredge
RD y volatile chemical emission during dredging
Rd b dissolved contaminant release for a clamshell dredge
Rj retardation factor as defined by Equation 53
Rp release rate of resuspended particles
Rp b particle resuspension rate
Ry.es volatile emission rate for chemical A from exposed sediment
Ry.esp volatile emission rate from partially filled vessel, g/cm2 sec
5 interphase contaminant transfer
Sc Schmidt number
A3
Appendix A Notation
-------
t
T
Tc
U
cb
V:
V,
W.
Ap
X
z
Z
A
a
7
7T
Pi
PS
Pw
Pb
V
Vl
T
Tcb
Subscript
1
2
3
time
temperature
dimensionless cycle time
Darcy or superficial water velocity
net deposition velocity
tangential velocity of cutterhead relative to axis of rotation
volume of the clamshell bucket
water velocity
cutterhead hydraulic inlet suction velocity
swing velocity of cutterhead dredge
maximum tangential velocity of cutterhead relative to fixed axis
wind speed
average pore water velocity
wind velocity
water current velocity
deposition velocity of sediment particles
total contaminant concentration in sediment (dry basis)
contaminant concentration sorbed to sediment
amount of substance adsorbed
distance through water, into the sediment or cap
water depth in meters
distance from top of vessel, cm
height of area swept by cutterhead as fraction of cutterhead
dimensions
length of area swept by cutterhead as fraction of cutterhead
dimensions
sediment porosity
air-filled porosity
Bohlen sweep area correction factor
3.14159...
air density
in situ bulk density of the sediment
water density
bulk density
kinematic viscosity
kinematic viscosity of air
dispersivity
clamshell bucket dredge cycle time
air
water
sediment
A4
Appendix A Notation
-------
Appendix B
A Priori Estimation of
Distribution Coefficients
Introduction
Application of methods for estimation of contaminant losses presented in
the main body of this report often requires estimation of equilibrium distribu-
tion coefficients between sediment, water, and/or air media. The distribution
coefficient is defined as the equilibrium ratio of the concentration of the con-
taminant in one phase divided by the concentration of the contaminant in an
adjacent phase at equilibrium.
Distribution coefficients represent the maximum amount of contaminant
that can be partitioned into an adjacent media given a concentration within the
sediment or dredged material solids. In general, distribution coefficients are
functions of temperature and concentration and the chemical properties of the
adjacent phases. By definition, however, distribution coefficients are not
functions of time nor the rate of mixing within the phases. Although the
coefficient is generally a function of concentration, the available data rarely
supports models that incorporate this behavior and the distribution coefficient
is typically assumed to be independent of contaminant concentration. Thus,
the equilibrium concentration in a phase is assumed to be linearly dependent
on the concentration in the adjacent phase, or for partitioning between water
and sediment solids (Equations 24, 25, and 26 of the main text)
c - c*
w Td
and for partitioning between water and air (Equations 31 and 32 of main text).
B1
Appendix B A Priori Estimation of Distribution Coefficients
-------
Of interest in this report is partitioning of contaminants between (a) sediment
or dredged material solids and pore water, and (b) sediments or dredged mate-
rial and adjacent air.
Contaminants of primary interest include metals and hydrophobic organic
chemicals that tend to partition strongly to sediments and thus pose long-term
sediment quality problems. Elemental species and hydrophobic organic mate-
rials partition to sediments by very different mechanisms. In addition, ele-
mental species tend to be nonvolatile, and their partitioning from sediments to
air or from water to air need not be considered. Thus, presentation of meth-
ods for the prediction of distribution coefficients of the materials of interest
will be separated by type of contaminant and the environmental interface
under consideration.
Hydrophobic Organic Species
Hydrophobic organic species are characterized by their low-water solubil-
ity. This class of compounds includes almost all hydrocarbons and substituted
organic compounds except the simple alcohols and phenols. Observations
have suggested that the partitioning of these compounds between water and a
particular soil or sediment is largely controlled by the hydrophobicity of the
compound, for example, as measured by the distribution coefficient of the
compound between water and octanol (Kow). In addition, observations have
suggested that the partitioning of a particular hydrophobic organic to different
soils or sediments is largely controlled by the organic carbon content of the
solid phase. This is consistent with the concept of "like dissolves like" for
defining the mechanism of sorption onto the sediment or soil phase. To a first
approximation, the distribution coefficient of a hydrophobic organic compound
between sediment or dredged material and water is given by
Kd = Kocfoc
where
Koc = organic carbon-based distribution coefficient
foc = fraction organic carbon in sediment or dredged material
The organic carbon-based distribution coefficient (Koc) is a measure of the
hydrophobicity of the organic compound. The fraction organic carbon is a
chemical property of the sediment or dredged material. This procedure for
estimating the distribution coefficient thus separates the problem into deter-
mining a single parameter characterizing the chemical and a single parameter
characterizing the sediment. The organic carbon-based distribution coefficient
is determined by measuring the sorption of a particular compound on a sedi-
ment or soil and normalizing by the organic carbon in the solid phase.
B2
Appendix B A Priori Estimation of Distribution Coefficients
-------
Lyman, Reehl, and Rosenblatt (1990) indicate that measured Koc values are
reasonably constant for a given compound sorbing to different soils and sedi-
ments. The coefficient of variation of Koc with different soils was 10 to
140 percent (Lyman, Reehl, and Rosenblatt 1990). Selected values of Koc are
included in Table Bl with values of other relevant physical properties. Data
from this table should be used with care recognizing that other data sources
might provide values that are orders of magnitude different. The values cho-
sen for the table, however, were selected based on consistency with similar
compounds and the availability of corroborating data, where possible.
Table B1
Physical Properties of Selected Compounds1
Compound
Acenaphthene
Aldrin
Anthracene
Benzolalpyrene
Chlordane
Chrysene
p,p'-DDT
Dieldrin
Fluoranthene
Fluorene
Hexachlorobenzene
Indenopyrene
PCB-1242 (Avg)
PCB-1254 (Avg)
Pentachlorophenol
Phenanthrene
Pyrene
TCDD
MW
154
365
178
252
410
228
354
381
202
166
285
276
261
327
266
178
202
322
f,
mm Hg
0.005
2.3Ex105
2 x 1 O'4
5.5 x10'9
1 x10'B
6.3 x10'9
1.9 x1Q-7
3 x 1 0'6
5x10'6
7 x 1 0'4
1 X10'5
1 x10'10?
4x10'4
7.7 x10'5
1.7 x10'4
6.8 x10'4
6.9 x10'7
1.4x1 O'9
Water
Solubility
mgll
3.47
0.017
0.045
0.004
0.056
0.002
0.003
0.2
0.26
1.6
0.005
0.062
0.24
0.057
20
1
0.135
0.0002
Hc
atm-m3
mol
2.9x10"*
6.5x10'4
0.001
4.6x1 0'7
9.6x10'6
9.5x1Q-7
3x10'5
7.5x1 0'6
5.1x10'6
9.7x10'5
7.5x10'4
5.9x10'10
5.7x10'4
5.8x10'4
3x1 0'6
1.6x10"*
1.4x10'6
3x10'6
L°g *oc
1.25
2.61
4.27
6
5.15
5.39
5.38
4.55
4.62
3.7
3.59
7.49
3.71
5.61
2.95
3.72
4.66
6.66
' From Groundwater Chemicals Desk Reference by Montgomery & Welkom, Lewis Publish-
ers (1990). Henry's Law Constants (H) calculated from vapor pressure and solubility.
A1 1 data are estimates at 25 °C.
Koc can also be estimated on the basis of correlations, for example with
solubility or the octanol-water partition coefficient (Kow). For example,
Curtis, Reinhard, and Roberts (1986) have presented the correlation
Appendix B A Priori Estimation of Distribution Coefficients
B3
-------
log Koc = 0.92 log Kow -0.23
As indicated by this correlation, Koc and Kow tend to be the same order of
magnitude. Kow is a good indicator of the ability of a compound to partition
between organic and water phases and has been correlated with bioconcentra-
tion and water solubility in addition to the sediment-water sorption coefficient.
In addition, Hansch and Leo (1979) have developed a procedure for estimating
Kow using only the molecular structure of the compound. A Kow for essen-
tially any compound whose structure is known can be estimated by this
method. Procedures and examples of various methods of estimating Kow, Koc,
and Kd are detailed in Lyman, Reehl, and Rosenblatt (1990).
The use of the approach discussed above is limited to situations where
sorption of the organic compound to the sediment is controlled by hydropho-
bic interactions. Hydrophilic compounds do not partition in the same manner
as the hydrophobic compounds. In addition, at very low organic carbon
contents, for example at 0.1 percent or less, direct sorption to mineral sur-
faces in the sediment or dredged material becomes important, and partitioning
is no longer simply a function of organic carbon content. Organic acids and
bases, phenolic compounds, and many pesticides can also deviate significantly
from the behavior suggested above at a pH that causes significant ionization of
the species. The acid dissociation constant (pKa) is a convenient indicator of
the extent of ionization. At a pH = pKa, half of the compound is in its ion-
ized state. At a pH = pKa + 2, the concentration of the ionized form is
100 times that of the concentration of the neutral species and the reverse is
true for pH = pKa - 2. Thus 2,4,6-trichlorophenol, with apKa = 7.42
(Montgomery and Welkom 1990), will interact hydrophobically with soils or
sediments and exhibit a Koc of about 1000 at pH < 6, while at pH > 9,
essentially no sorption will be observed.
In addition to the limitations outlined above, the assumption of constant Kd
typically limits the validity of the entire approach to low-contaminant concen-
trations. A critical sediment loading can be defined as the sediment concen-
tration that is in equilibrium with a saturated water solution, i.e., water
containing the compound at its solubility limit. Linear partitioning is typically
not observed at sediment concentrations that are near the critical loading. In
addition, under no circumstances should linear partitioning between sediment
and water be applied at sediment concentrations that exceed the critical load-
ing. It is possible to measure a sediment concentration that exceeds critical
loading due to the presence of a separate nonaqueous phase or due to nonlin-
ear partitioning. It is not possible, however, to achieve a truly dissolved
concentration that exceeds the water solubility of the compound.
At low-sediment concentrations, for example, in sediment suspended in the
water column, linear partitioning is also apparently no longer observed; distri-
bution coefficients depend on sediment concentration, perhaps due to the
presence of colloidal material (Gschwend and Wu 1985). Baker et al. (1991)
observed that field data on partitioning to dilute suspended solids rarely fit the
64
Appendix B A Priori Estimation of Distribution Coefficients
-------
linear partitioning model, perhaps due to the presence of colloids and the
kinetics of solids uptake of the sorbing contaminant.
Elemental Species
The partitioning of metals and other elemental species between sediments
or dredged material and water is much more complicated than that for hydro-
phobic organic species. The total concentration of an element in sediment or
dredged material is the sum of that which is chemical bound in various geo-
chemical phases (Brannon et al. 1976), physically sorbed, and dissolved in the
interstitial waters. Generally, the chemically bound portion, which usually
comprises 90 to 99 percent of the contaminant mass, is immobile and
unavailable for partitioning into the aqueous phase under most environmental
conditions. The physically sorbed portion is exchangeable through ion
exchange, and the dissolved form is mobile. In reality, the elemental species
exist in the sediment in a variety of forms. Brannon et al. (1976) identified at
least the following sediment geochemical phases for sorbed and fixed
contaminants:
a. Adsorbed on the surface of charged mineral and organic surfaces.
b. Oxides, hydroxides, and hydrous oxides of Mn and Fe.
c. Chemically bound in organic matter.
d. Chemically bound with sulfides.
e. Bound within the crystalline lattice (residual).
Brannon et al. (1976) devised a selective extraction scheme that treated
sediment samples with increasingly harsh treatments to define the interstitial
water, exchangeable, easily reducible, organic + sulfide, moderately reducible
and residual fractions. The contaminants removed with each fraction were
assumed to indicate the proportion of the original element in each of the
chemical forms identified above. The exchangeable fraction, for example,
was defined by the amount of the elemental species that could be extracted
with ammonium acetate. Brannon et al. (1976) applied the selective extraction
procedure to sediments from three areas, Ashtabula, Ohio (freshwater),
Mobile Bay, Alabama (estuarine), and Bridgeport, Connecticut (saltwater).
The exchangeable fraction of iron and manganese was found to correlate well
with the interstitial water concentrations with an exchangeable fraction-water
distribution coefficient of about 9 i/kg for both.
Zinc and nickel correlated less well with the exchangeable concentration
and exhibited an average exchangeable fraction-water distribution coefficient
of 9 and 5, respectively. Copper and cadmium were not found in detectable
quantities in the exchangeable fraction and interstitial water; concentrations
were 100 to 1,000 times lower than for the other species. Thus, the
B5
Appendix B A Priori Estimation of Distribution Coefficients
-------
exchangeable fraction, as defined by the amount of contaminant extracted with
ammonium acetate, was a reasonable indicator of the interstitial or presumed
equilibrium water concentrations. This would assume no changes in the
chemical state of the sediment. Oxidation of the sediment, for example, tends
to mobilize many of the metals and other elemental species that might be
present in the sediment.
Brannon, Myers, and Price (1992) and Environmental Laboratory (1987)
conducted further tests to define distribution coefficients for elemental species
in freshwater sediments at Indiana Harbor, Indiana, and Hamlet City Lake,
North Carolina. Both sequential batch leach tests and continuous column
leaching tests were employed. The combination of batch and continuous tests
has several advantages over traditional procedures to determine mobile ele-
mental fractions. The batch test ostensibly defines equilibrium conditions for
a particular sediment to water ratio, while the continuous test should indicate
the dynamics of the leaching process. However, further research is required
to fully define the capabilities and procedures for conducting and analyzing
batch and column leach tests.
Although definitive procedures for the a priori estimation of elemental
distribution coefficients do not exist, general guidelines for the magnitude of
these distributions coefficients do exist. The elemental partitioning studies
(Brannon et al. 1976) and the batch equilibrium and column leaching studies
(Environmental Laboratory 1987; Brannon, Myers, and Price 1992) indicate
that distribution coefficients for most metals in freshwater sediments range
from 1 to 10 I /kg. Table B2 lists the observed range of distribution coeffi-
cients adapted from a table presented by Dragun (1988). The distribution
coefficients summarized in Table B2 are the ratio of the total soil or sediment
concentration (i.e., the sum of both exchangeable and chemically fixed ele-
ments) to the adjacent water concentrations. As indicated by the discussion
above, it is believed that a more generally useful partition coefficient would be
one based on the exchangeable concentration of the element.
Air-Water Partitioning
Essentially all of the elemental species of interest in sediments and dredged
material are nonvolatile. Therefore, the discussion here will be limited to
evaporation of hydrophobic organic species from water.
The general expression of equilibrium at a fluid-fluid interface is based on
the concept of continuity of component activity across the interface. This is
most easily expressed as continuity of fugacity, which is an effective pressure
corrected for nonidealities.
The fugacity on each side of the interface for a contaminant i is written as
the product of a standard state fugacity, the mole fraction of the contaminant
and a correction for nonideality, the activity coefficient.
Rfi
Appendix B A Priori Estimation of Distribution Coefficients
-------
Table B2
Ranges for Distribution Coefficients for Various Soils and Clays
(After Dragun 1988)
Element
Ag
Am
As(ill)
As(V)
Ca
Cd
Ce
Cm
Co
Cr(lll)
Cr(VI)
Cs
Cu
Fe
K
Mg
Mn
Mo
Np
Pb
Po
Pu
Ru
Se(IV)
Sr
Tc
Th
U
Zn
Observed Kd
ml/g
10-1,000
1-47,230
1-8.3
1.9-18
1.2-9.8
1.3-27
58-6,000
93-51,900
0.2-3,800
470-150,000
1.2-1,800
10-52,000
1 .4-333
1.4-1,000
2-9
1.6-13.5
0.2-10,000
0.4-400
0.2-929
4.5-7,640
196-1,063
11-300,000
48-1,000
1.2-8.6
0.2-3,300
0.003-0.28
2,000-510,000
11-4,400
0.1-8,000
Logarithmic Mean
110
810
3.3
6.7
4
6.7
1,100
3,300
55
2,200
37
1,100
22
55
5.5
5.5
148
20
11
100
550
1,800
600
2.7
27
30
60,000
45
16
Standard Deviation
(Error Factor)1
3.7
20
1.8
1.6
2.2
2.5
3.7
6.7
10
3.3
9
6.7
3
5.5
1.6
1.6
15
8.2
10
5.5 ||
2 U
10 I
2.7
2
7.4
3
4.5
3.7
6.7
1 Standard deviation as a multiplicative factor of mean.
Appendix B A Priori Estimation of Distribution Coefficients
B7
-------
where
/• = fugacity of component i
Xj = mole fraction of i
7, = activity coefficient of component i
fj° = standard state fugacity of component i
At the air-water interface, the equality of fugacities implies
*air _ ,.water
Ji ~ Ji
The relationship between concentrations (or mole fractions) of the contaminant
across the air-water interface then depends on the specification of activity
coefficients and standard state fugacities in each of the phases.
The standard state fugacity is normally taken as the pressure that would be
exerted by the pure component (i.e., contaminant i) at the same temperature,
pressure, and phase as the mixture. Thus the standard state fugacity of a
component in air would be the pressure exerted by a pure component vapor in
the atmosphere (i.e., 1 atm). In addition, since gases act ideally at low pres-
sure, the activity coefficient in the atmosphere is 1. Similarly, the standard
state fugacity of a component in water would be the pressure exerted by a
pure component liquid at the desired temperature. This is just the pure com-
ponent vapor pressure (or saturation pressure) at that temperature. Estimation
of the activity coefficient in water is more difficult due to the typically large
deviations from ideality (i.e., y"T f 1). Hydrophobic organics exhibit a low
solubility in water, and even a saturated water solution is not changed appre-
ciably by the presence of the organic species. Thus, the water-organic
interactions and the activity coefficient are essentially independent of concen-
tration. In addition, a saturated solution exerts the same component pressure
as a pure phase since the addition of any more of the component produces
such a phase. Thus, continuity of fugacities across the air-water interface for
a saturated water solution is described by
y;.(DP = PV = x,,7,pv
or
B8
Appendix B A Priori Estimation of Distribution Coefficients
-------
where
yt = mole fraction of i in the air (ytP is the component partial pressure)
P = total pressure (1 atm)
Pv = pure component vapor pressure of i (= yf for saturated air)
xis = mole fraction of i in water at solubility limit
7, = activity coefficient of i in water
Based on this approach, the relationship between air concentration (as mea-
sured by air phase mole fraction, yt and water concentration (as measured by
water phase mole fraction, x,) is given by
Xis
= Hx,
where
H = P/J^, a Henry's Law Constant
Thus, the air-water equilibrium for a hydrophobic organic compound is
also governed by a linear partitioning law with an essentially constant distribu-
tion coefficient. This is valid as long as the water phase activity coefficient is
independent of the concentration of the partitioning contaminant, an assump-
tion that is generally good for low solubility, hydrophobic organic com-
pounds. This approach cannot be applied to hydrophilic organic compounds
such as phenols, low molecular weight alcohols, and organic acids or bases.
The Henry's Law Constant, H, is defined above as the ratio of a vapor
pressure and the solubility in mole fraction units. It is often convenient to
define solubility in concentration, or mass per volume units. The equivalent
Henry's Law Constant (Hc) is the ratio of the pure component vapor pressure
to the solubility in these concentration units. The relationship between the
partial pressure in the air (y,P) and the water concentration then becomes
Y,P = Hect
Henry's Law Constants reported in Table Bl are the Hc as defined here with
atmospheres used as the unit of pressure and concentration measured in
mole/cubic meter. The values for Hc shown in Table Bl were calculated from
B9
Appendix B A Priori Estimation of Distribution Coefficients
-------
vapor pressure and solubility data. This approach was taken because several
independent measurements of solubility and vapor pressure are reported by
Montgomery and Welkom (1990), whereas typically only a single Henry's
Law Constant value is reported. It is also convenient at times to use concen-
tration units in the air phase. The Henry's Law Constant is then dimension-
less if the same mass and volume units are used to define the concentrations in
the air and water phases. Note also that although H was used to indicate unit
of pressure/mole fraction and Hc units of pressure/concentration, this notation
is not standardized. Errors have often resulted from the use of incorrect units
for Henry's Law Constants, especially when this quantity is used in mass
transfer two-layer resistance models.
Partitioning Between Sediments Or Dredged
Material and Air
The partitioning of hydrophobic organics between exposed sediments or
dredged material and air is generally equivalent to that defined in the previous
section. As long as the volumetric water content of the sediments is more
than a few percent, water will tend to coat the surface of the sediment parti-
cles. Phase equilibrium is then controlled by the water-air interface. The
equilibrium interstitial water concentration as defined by the sediment-water
equilibrium can then be used with a Henry's Law Constant as defined by the
preceding section to determine the equilibrium partial pressure of the contami-
nant above the exposed sediment.
When the volumetric water content of the sediment or dredged material is
less than a few percent, the equilibrium partial pressure of hydrophobic organ-
ics of the surface of the sediment begins to decrease dramatically. Direct
sorption of the organic molecules onto the sediment surface can take place, a
process which can significantly increase the capacity of the solid phase to
retain contaminants. The partitioning between the air and sediment phase is
largely a function of the exposed surface area of the sediment phase. Valasaraj
and Thibodeaux (1992) have presented data on a number of hydrophobic
organic compounds sorbing to dry, moist, and wet soils. On completely dry
soils or sediments, sorption of vapors onto the soil surface is apparently con-
trolled by the surface area of the sorbet and is a nonlinear function of concen-
tration. In general, however, sediments exposed due to tides or dredged
material in a confined disposal facility will rarely achieve the dry state neces-
sary for this mechanism to become important. Even dryland soils are unlikely
to be influenced by this process except in the upper few centimeters of soil.
In addition, the assumption of water-wet sediment will provide a conservative
upper bound to the equilibrium partial pressure of a contaminant above a dried
sediment. For these reasons, equilibrium at a dry sediment-air interface will
not be considered further.
B10
Appendix B A Priori Estimation of Distribution Coefficients
-------
References
Baker, J. E., Eisenreich, S. J., and Swackhamer, D. L. (1991). "Field mea-
sured associations between polychlorinated biphenyls and suspended solids in
natural waters: An evaluation of the partitioning paradigm." Organic sub-
stances and sediments in water. R. A. Baker, ed., Lewis Publishers.
Brannon, J. M, Engler, R. M., Rose, J. R., Hunt, P. G., and Smith, I.
(1976). "Selective analytical partitioning of sediments to evaluate potential
mobility of chemical constituents during dredging and disposal operations,"
Technical Report D-76-7, U.S. Army Engineer Waterways Experiment Sta-
tion, Vicksburg, MS.
Brannon, J. M., Myers, T. E., and Price, C. B. (1992). "Leachate testing
of Hamlet City Lake sediment," Miscellaneous Paper D-92-5, U.S. Army
Engineer Waterways Experiment Station, Vicksburg, MS.
Curtis, G. P., Reinhard, M., and Roberts, P. V. (1986). ACS Symposium
Series, 323.
Dragun, J. (1988). The soil chemistry of hazardous materials. Hazardous
Materials Control Research Institute, Silver Spring, MD.
Environmental Laboratory. (1987). "Disposal alternatives for PCB-
contaminated sediments from Indiana Harbor, Indiana," Miscellaneous Paper
EL-87-9, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
MS.
Gschwend, P. M., and Wu, S. (1985). "On the constancy of sediment-
water partition coefficients of hydrophobic pollutants," Environmental Science
and Technology 19, 90-96.
Hansch, C., and Leo, A. J. (1979). Substituent constants for correlation
analysis in chemistry and biology. John Wiley, New York.
Lyman, W. J., Reehl, W. F., and Rosenblatt, D. H. (1990). Handbook of
chemical property estimation methods. American Chemical Society, Washing-
ton, DC.
Montgomery, J. H., and Welkom, L. M. (1990). Ground-water chemicals
desk reference. Lewis Publishers, Chelsea, MI.
Valsaraj, K. T., and Thibodeaux, L. J. (1992). "Equilibrium adsorption of
chemical vapors onto surface soils: Model predictions and experimental
data." Fate of pesticides and chemicals in the environment. J. L. Schnoor,
ed., John Wiley, New York.
B1 1
Appendix B A Priori Estimation of Distribution Coefficients
-------
Appendix C
Input Parameters for Disposal
From an Instantaneous Dump
(DIFID), Disposal From
Continuous Discharge (DIFCD),
and Disposal From a Hopper
Dredge (DIFHD) Models
Table C1
Model Input Parameters
Parameter
Models1
Units
Options2
Disposal Site Descriptions
Descriptive title
Gridpoints (left to right)
Gridpoints (top to bottom)
Distance between gridpoints
Constant water depth
Gridpoints depths
Points in density profile
Depth of density point
Density at profile point
Bottom slope in x-direction
Bottom slope in z-direction
Site boundary grid locattons
,C,H
,C,H
,C,H
,C,H
,C,H
,C,H
,C,H
,C,H
,C,H
,H
,H
,C,H
feet
feet
feet
feet
g/cc
degrees
degrees
C
V
(Sheet 1 of 4)
1 The use of a parameter in the DIFID, DIFCD, and DIFHD models is indicated in the table
by an 1, C, or H, respectively.
2 The use of a parameter for the constant depth option or variable depth option is indi-
cated in the table by a C or V, respectively. Other optional uses for parameters are so
indicated.
Appendix C Input Parameters
C1
-------
Table C1 (Continued)
Parameter
Models1
Units
Options2
Disposal Operation Descriptions
Volume of material in barge
Discharge flow rate
Radius of discharge
Discharge depth
Angle of discharge
Dredge course
Vessel speed
Barge velocity in x-direction
Barge velocity in z-direction
Barge length
Barge width
Post-disposal depth
Bottom depression length in x-direction
Bottom depression length in z-direction
Bottom depression depth
X-coordinate of disposal operation
Z-coordinate of disposal operation
Disposal duration
Time from start of tidal cycle
Number of hopper bins opening together
Distance between bins
I
C,H
C,H
C,H
C
C
C
I
I
I
I
I
I,H
I,H
I,H
I.C.H
I,C,H
I,C,H
I,C,H
H
H
cu yd
cfs
feet
feet
degrees
degrees
ft/sec
ft/sec
ft/sec
feet
feet
feet
feet
feet
feet
feet
feet
seconds
seconds
feet
Optional
Optional
Optional
Disposal Site Velocity Descriptions
Type of velocity profile
Tidal cycle time of velocity if constant
profile not used
Vertically averaged velocity in x-direction
at gridpomts
Vertically averaged velocity in z-direction
at gridpoints
Velocity in x-direction at upper point
Depth of upper point for x-direction
velocity
Velocity in x-direction at lower point
Depth of lower point for x-direction
velocity
Velocity in z-direction at upper point
Depth of upper point for z-direction
velocity
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
seconds
ft/sec
ft/sec
ft/sec
feet
ft/sec
feet
ft/sec
feet
V
V
V
C
C
C
C
C
C
(Sheet 2 of 41
C2
Appendix C Input Parameters
-------
Table C1 (Continued)
Parameter | Models1
Units
Options2
Disposal Site Velocity Descriptions (Continued)
Velocity in z-direction at lower point
Depth of lower point for z-direction
velocity
I,C,H
I,C,H
ft/sec
feet
C
C
Material Descriptions
Water density at dredge site
Number of solid fractions
Solid fraction descriptions
Solid fraction specific gravity
Solid fraction volumetric concentration
Solid fraction settling velocity
Solid fraction deposited void ratio
Moisture content of material in barge as
multiple of liquid limit
Bulk density of dredged material
Liquid phase contaminant concentration
Background contaminant concentration
Sediment contaminant concentration
Contaminant water quality criteria
Toxicity criteria
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
1
I,C,H
I,C,H
I,C,H
I.C.H
I,C,H
I,C,H
g/cc
cu ft/cu ft
ft/sec
g/cc
mg/f
mg/f
mg/kg
mg/f
percent
Cohesive
Optional
Optional
Optional
Optional
Optional
Model Coefficients
Settling coefficient
Apparent mass coefficient
Drag coefficient
Form drag for collapsing cloud
Skin friction for collapsing cloud
Drag for an ellipsoidal wedge
Drag for a plate
Friction between cloud and bottom
Horizontal diffusion dissipation
Vertical diffusion coefficient
Cloud/ambient density gradient ratio
Turbulent thermal entrainment
Entrainment in collapse
Jet entrainment
Thermal entrainment
Entrainment by convection in collapse
Entrainment due collapse of element
I,C,H
I,C,H
I,C,H •
I.C.H
I,C,H
I,C,H
I.C.H
I.C.H
I,C,H
I.C.H
I,C,H
I,H
I,H
H,C
H,C
C
C
(Sheet 3 of 41
Appendix C Input Parameters
C3
-------
Table C1 (Concluded)
Parameter
Models1
Units
Options2
Input, Output, and Execution Descriptions
Processes to simulate
Type of computations to perform for
initial mixing
Number of depths for initial mixing
calculations
Depths for initial mixing calculations
Duration of simulation
Time steps for mixing calculations
Convective descent output option
Collapse phase output option
Number of print times for initial mixing
output
I,C,H
I,C,H
I,C,H
I,C,H
I,C,H
1,C,H
I,C,H
I,C,H
I,C,H
feet
seconds
f Sheet 4 of 4)
C4
Appendix C Input Parameters
-------