vvEPA
United States
Environmental Protection
Agency
Office of
Radiation Programs
Washington, D.C. 20460
EPA 520/1 -86-016
September 1988
Radiation
Prediction of
Transport of Low-Level
Radioactive Middlesex Soil
at a Deep-Ocean Disposal
Site
-------
-------
PREDICTION OF VERTICAL TRANSPORT OF LOW-LEVEL RADIOACTIVE
MIDDLESEX SOIL AT A DEEP-OCEAN DISPOSAL SITE
by
James S. Bonner
Applied Technology Division
Computer Sciences Corporation
EPA Environmental Research Laboratory
Narragansett, Rhode Island 02882
Carlton D. Hunt
Marine Ecosystems Research Laboratory
Graduate School of Oceanography
University of Rhode Island
Narragansett, Rhode Island 02882
John F. Paul and Victor J. Bierman, Jr.
Environmental Research Laboratory
Office of Environmental Processes and Effects Research
U.S. Environmental Protection Agency
South Ferry Road
Narragansett, Rhode Island 02882
Prepared for
Analysis and Support Division (ANR-461C)
Office of Radiation Programs
U.S. Environmental Protection Agency
401 M Street, S.W.
Washington, D.C. 20460
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
NARRAGANSETT, RHODE ISLAND 02882
-------
DISCLAIMER
This report has been reviewed by the Environmental Research
Laboratory, U.S. Environmental Protection Agency, Narragansett, Rhode
Island, and approved for publication. Approval does not signify that
the contents necesarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of trade-names
or commercial products constitute endorsement or recommendation for
use.
-------
FOREWORD
In response to the mandate of Public Law 92-532, the
Marine Protection, Research and Sanctuaries Act of 1972, as
amended, the Environmental Protection Agency (EPA) has
developed a program to promulgate regulations and criteria to
control the ocean disposal of radioactive wastes. Interest
expressed by other Government agencies has led EPA to consider
the environmental fate of unpackaged soils, containing very low
levels of naturally-occurring radioactivity, after surface
disposal over a deep-sea site. An important technical factor
in any environmental assessment of this disposal alternative
for dry soils is the potential for physical transport of the
material from the point of initial disposal at the ocean
surface.
This report summarizes the data obtained through
laboratory experiments using both small and large experimental
tanks to simulate such a disposal operation. A model was then
developed to predict deep-ocean transport. Model calculations
were made to incorporate soil settling rates and physical and
chemical properties which could influence the behavior of
contaminants in the soil on initial contact with seawater and
subsequent passage through the water column. The study was
carried out in a manner consistent with the Hazard Assessment
Strategy developed by the EPA Office of Water. The Hazard
Assessment Strategy contains the scientific framework under
which potential ocean disposal permit requests may be evaluated.
The experimental methods, experimental results and modeling
approach and application to predict deep-ocean transport are
described, and conclusions drawn regarding fate of the soils if
they were to be disposed of at the ocean surface over a 4,000
meter depth site.
The Office of Radiation Programs will use this report as
an information base for any future inquiries regarding the
ocean disposal of soils containing low levels of radio-
activity. The methodologies described and the model developed
may also be valuable for the environmental assessments of other
kinds of pollutants proposed for ocean disposal in unpackaged
form.
The Agency invites all readers of this report to send any
comments or suggestions to Mr. David E. Janes, Director,
Analysis and Support Division, Office of Radiation Programs
(ANR-461), Environmental Protection Agency, Washington, D.G.
20460.
Sheldon Meyers,Director
Office of Radiation Programs (ANR-458)
-------
ABSTRACT
Potential ocean disposal of low-level natural radioisotopes
associated with soils was investigated by combining experimental and
modeling approaches to determine transport and fate of the material.
The work was accomplished within the framework of the hazard assessment
methodology developed at the EPA Environmental Research Laboratory
Narragansett. Source material was provided from a Department of Energy
designated Formerly Utilized Sites Remedial Action Program (FUSRAP)
site in Middlesex, New Jersey. The experimental approach involved
characterization of the source material for particle size distribution,
specific gravity, total radioisotope activity and distribution,
nuclide soluble phase equilibria of the radioisotopes, and particle
settling velocities.
Soil particles were primarily sandy with a size range from less
than 63 microns to greater than 2000 microns and a specific gravity of
2.31. The median particle size for total gamma plus beta or total
alpha counts (125 microns) was not coincident with the median grain
size (350 microns). The individual radioisotopes Ra-226, U-234,
U-238, and Th-230 exhibited similar distributions (median size of 250
microns), while the distributions for Pb-210 and Po-210 were shifted
to the larger particle sizes. Radioisotopes were primarily associated
with discrete soil particles and are assumed to be residuals of the
original ores processed at the site. Less than 10 per cent of the
associated radioisotopes leached from the soil after exposure to
seawater for up to 20 hours. Particle settling velocities measured
for a number of size classes in a 1 meter settling column ranged up to
8.2 cm/sec (median 2.1 cm/sec). Mesocosm-scale experiments confirmed
that settling would be the dominant vertical transport mechanism for
these soil particles at a deep-ocean disposal site.
The experimental results were used to calibrate a one-dimensional
convective-diffusive particle transport model which was applied to a
hypothetical ocean disposal site in 4000 m of water. The model
predicted that 95 per cent of the soils and associated radioisotopes
would impact the bottom sediments within 4.5 days. Addition of a
horizontal transport component to the model indicated that 95 per cent
of the soil mass would impact the bottom sediments within 40 km of the
disposal point along the direction of mean flow, for typical currents
observed off the northeast U.S. continental shelf.
An overall project with the FUSRAP material was conducted which
involved physical-chemical (exposure) and ecosystem (fate/effects)
components. The physical-chemical component is described in the work
reported here. The ecosystem component is covered in a separate
report.
This report covers a period from 1 January 1984 to 30 April 1985,
and work was completed as of 30 April 1985.
IV
-------
CONTENTS
Page
Foreword iii
Abstract iv
Figures ...vi
Tables vii
Acknowledgements viii
1. Introduction 1
Background 1
Scope 1
2. Experimental Methods 5
Soil Samples 5
Specific Gravity 5
Particle Characteristics 5
Activity Distributions 6
Soluble Phase Equilibria 6
Settling Velocity Distributions 6
Mesocosm-Scale Experiments 8
3. Experimental Results 11
Size Distribution 11
Activity Distributions 11
Isotope Dissolution 22
Settling Velocity Distributions 22
Mesocosm-Scale Experiments 31
4. Modeling Approach 34
Model Development 34
Level 1 34
Level 2 ,.. 36
Model Calibration and Data Synthesis 37
Model Application 37
5. Conclusions 42
References 44
Appendices
A. Bench-Scale Settling Velocity Data.... 45
B. Mesocosm-Scale Data 50
v
-------
FIGURES
Number Page
1 Hazard assessment strategy for ocean disposal 2
2 Laboratory settling column 7
3 MERL mesocosm 9
4 Cumulative mass versus particle size 12
5 Size distribution between 20 and 600 microns as
determined by wet and dry methods 13
6 Size distributions of particles separated into size
classes by dry sieving, as determined by wet and
dry methods 14
7 Cumulative gross activity versus particle size. 19
8 Cumulative individual isotope activity versus
particle size 21
9 Settling velocity distribution for Middlesex soil
determined from laboratory column experiments....... 25
10 Settling velocity distributions for Middlesex soil
separated into size classes by dry sieving 26
11 Mass removal experimental results for particles
greater than 63 microns and with 11 cm^/sec
dispersion in MERL mesocosm. 32
12 Dye experimental results with 11 cm^/sec dispersion
in MERL mesocosm 38
13 Time of arrival at the bottom of 4000 meter water
column for Middlesex soil 39
14 Bottom impact distance along mean flow direction for
Middlesex soil disposed in 4000 meters of water....
41
VI
-------
TABLES
Number
1
2
Radioisotope activity for Middlesex soil,
Page
20
Isotope solubility and partition coefficients for
Middlesex soil 23
Mass balance estimates of isotope solubility for
Middlesex soil from MERL mesocosm experiments 24
Effects of dispersion on mass removal of Middlesex
soil from MERL mesocosm experiments 33
VI1
-------
ACKNOWLEDGEMENTS
We thank Henry A. Walker, Robert R. Payne, J. Neiheisel, and
S.L. Kupferman for their critical comments on this manuscript. We
would like to acknowledge the efforts of Jan C. Prager during the
settling column construction and for providing help during the
measurement of the settling velocity distributions. Fred Russel
provided technical assistance in the mesocosia-scale experiments.
Colette Brown assisted in the preparation of this report. Radioisotope
counting was performed by the Eastern Environmental Radiation Facilities,
U.S. Environmental Protection Agency. This work was supported in
part by a grant to the Marine Ecosystem Research Laboratory of the
University of Rhode Island from the U.S. EPA Office of Radiation
Programs, Marilyn Varela, Project Officer. Contribution No. 721 of
the Environmental Research Laboratory - Narragansett.
viii
-------
SECTION 1
INTRODUCTION
BACKGROUND
The Marine Protection, Research and Sanctuaries Act of 1972 (33
U. S. C. 1401), commonly referred to as the Ocean Dumping Act,
authorizes the Environmental Protection Agency (EPA) to regulate all
ocean disposal activities in the United States, including disposal of
radioactive wastes not specifically prohibited by law. Under the
provisions of this Act, EPA is also required to establish and apply
criteria for review and evaluation of disposal permit applications.
In 1981 Sandia National Laboratories initiated an evaluation of
potential disposal sites, including the oceans, for soils contaminated
with low, but significant, levels of natural radioisotopes under the
Department of Energy's Formerly Utilized Sites Remedial Action Program
(FUSRAP). This program was designed to evaluate and clean up numerous
'inactive industrial plant sites remaining from the Manhattan Project
(Kupferman et al., 1984). Sandia identified the sampling plant
material at Middlesex, New Jersey, as a candidate for potential ocean
disposal. Efforts were begun to document the feasibility and
advantages of ocean disposal of the soils and rubble from this site
and to support potential permit applications to EPA (Kupferman et al.,
1984).
In order to assist the EPA Office of Radiation Programs in
evaluating potential permit applications of this type, the EPA
Environmental Research Laboratory - Narragansett (ERLN) proposed that
a hazard or risk assessment methodology (Prager et al., 1984) be
developed specific to disposal of low-level radioactive waste in the
oceans. The principal objective of this effort was to develop a
useful, scientifically credible framework within which the potential
permitting process could be undertaken. This methodology required
several distinct types of effort including waste characterization,
site selection, and determination of exposure fields and potential
biological impacts at any of several levels of sophistication (Figure
1).
Preliminary characterization of the site in Middlesex by Sandia
had identified several potential contaminants including natural
radionuclides and inorganic and organic toxins (Kupferman et al.,
1984). Of these, only radioisotopes were at levels to be of potential
concern. In addition, Sandia identified the 106-Mile Ocean Disposal
Site, a previously designated site for industrial waste disposal in
the northwest Atlantic off of the continental shelf, as a candidate
site to receive the soils (Kupferman et al., 1984).
SCOPE
The Sandia studies and results were useful but not sufficient to
develop and test the hazard assessment methodology. This methodology
-------
SITE
CHARACTERIZATION
HASTE
CHARACTERIZATION
EXPOSURE
ASSESSMENT
LEVEL 1
LEVEL N
HAZARD
ASSESSMENT
DISPOSAL
DECISION
MONITORING
VALIDATE/REVISE
HAZARD ASSESSMENTS
EFFECTS
ASSESSMENT
LEVEL 1
LEVEL N
Figure 1. Hazard assessment strategy for ocean disposal.
-------
specificially required better understanding of the exposure fields
which would result from disposal of this soil in the marine
environment. Since field measurements are difficult to obtain without
an actual disposal operation, it was determined that a
laboratory-scale experimental .effort be combined with a modeling
approach to provide the necessary transport information. The
availability of soils from a designated FUSRAP, now FODOCS
(Feasibility of Disposal of Contaminated Soils), site allowed us to
conduct experiments with a specific low-level radioactive material,
within the context of the hazard assessment methodology.
Development of the model required estimates of soil settling
rates and physical and chemical properties that would modify the
contaminants in the soil on contact and passage through sea water.
Specifically, it was necessary to:
1. Characterize the Middlesex soil for parameters affecting
transport.
2. Develop and calibrate a vertical transport model using data
from 1.
3. Apply this model to estimate the transport and fate of the
radioisotopes at a hypothetical deep-ocean disposal site.
The experimental efforts necessary for this development included:
1. Measurement of particle specific gravity, size and activity
distributions, and radionuclide dissolution for
representative soils from the Middlesex site.
2. Construction of a laboratory bench scale settling column and
measurement of dynamic soil settling velocity distributions
under quiescent hydrodynamic conditions.
3. Simulation of deep-ocean vertical dispersion conditions and
quantification of vertical dispersion over a range of ocean
values by modification of large scale water columns
(mesocosms) at the Marine Ecosystems Research Laboratory
(MERL), University of Rhode Island.
4. Measurement of vertical transport of Middlesex soil in the
mesocosms under vertical dispersion conditions typical of the
deep ocean.
The model used to describe transport of this soil in the ocean
depended on the characteristics of the soil and behavior of associated
contaminants. Tightly bound contaminants required a description of
particle transport alone to determine the contaminant exposure
concentration fields. If, however, a significant fraction of the
contaminant had not remained associated with the particulate matrix,
-------
then the modeling approach would also have had to consider a soluble
contaminant phase. In either case, radioisotope distributions for the
source material were required to predict transport and fate,
particularly if the activity was unevenly distributed throughout the
particle size spectrum. In this case, model development would have
focused on the transport of specific particle size classes which had
the significant levels of radioactivity.,
Our efforts demonstrated that a particulate convective-diffusive
model was sufficient to describe the transport and fate of natural
radionuclides associated with soil from the FUSRAP site. This model
predicted that the vast majority of soils would impact the bottom
sediments at a 4000 m disposal site within 4.5 days and up to 40 km
away, along the direction of mean flow for typical currents observed
off the northeast U.S. continental shelf.
-------
SECTION 2
EXPERIMENTAL METHODS
SOIL SAMPLES
Representative samples of the soil from the Middlesex site were
obtained in August 1982 by Sandia and MERL personnel. Samples of soil
were excavated from the site, sieved through a 1/4 inch (ca. 6350 micron)
mesh, characterized on site for total radioactivity, and combined into five
separate 100 kg subsamples having different total radioactivity.
Samples were homogenized at the site in a 0.15 m^ standard cement
mixer and divided into 50 kg replicate batches. These samples were
distributed by Sandia for acute solid phase toxicity testing and an
experiment involving the controlled marine ecosystem concept at MERL.
One replicate of each sample was archived at Middlesex for possible
future work. The results in this report are based on the sample sets
Designated B1A and B) which had the highest activity (ca. 500 pCi
Ra-226/g dry weight). Specific activity determinations were carried
out in 1982 by Sandia and in 1984 by the Eastern Environmental
Radiation Facilities (EERF) of EPA in Montgomery, Alabama. The
activity in this sample is about 10 times higher than the mean for the
total Middlesex site (Kupferman, personal communications).
SPECIFIC GRAVITY
Specific gravity of the Middlesex soil sample was determined
using standard pycnometric procedures at 20 degrees centigrade. The
average bulk specific gravity was 2.31. Throughout this study, this
value was assumed for all size categories.
PARTICLE CHARACTERISTICS
Soil size distributions were measured by two methods. Method one
involved quantitatively oven drying 100 g of soil at 70°C to determine
percent moisture. The dried sample was then sieved for 20 minutes
with a Ro-tap® particle sieve shaking system (Carver, 1971) using 0.5
Phi size increments between 4.0 and - 1.0 Phi units (63 to 2000
microns). The soil retained on each sieve was quantified in terms of
mass and the mass percent, and a cumulative size distribution
determined.
The second technique utilized a HIAC-SSTA automatic particle
counter calibrated to count particles in 5 sizes between 20 and 600
microns. The counting sensor was positioned vertically at the bottom
of a 1 m bench scale experimental settling column (discussed later),
filled with 0.4 micron filtered.lower Narragansett Bay sea water
(salinity ca. 30 ppt). A known mass of soil was added at the top of
the column and allowed to settle through it. Total particle counts
were maintained at less than 100/ml to avoid coincidence counts with
the sensor. The hydrostatic head from the column was sufficient to
force the fluid through the sensor at 380 ml/min. Bias resulting from
-------
differential particle arrival time was eliminated by continuing the
experiment until background count rates were obtained. This method
provided a measure of the dynamic behavior of the wetted particles.
Particle counts were transformed to mass by multiplying average
particle volume (assuming spherical particles) for a given size range
by the measured average soil density. Comparison of size classes
between the two techniques was accomplished by adjusting and
normalizing the dry size results to correspond with the HIAC results.
ACTIVITY DISTRIBUTIONS
Particles obtained from the dry size classification procedure
were analyzed for total radioactivity by two techniques. A gross
measure of total activity was determined on each size class by
counting total gamma plus beta and total alpha particles emanating
from a known mass of sample with a Nuclear Measurements, Inc., PC-5
proportional counting system. Results were calculated in terms of
disintegration per minute per gram dry sample (DPM/g) and corrected for
incident background radiation. These same samples were then sent to
EERF for quantification of radioisotopes Ra-226, Pb-210, Po-210,
U-234, U-235, U-238, Th-227, Th-228, Th-230, and Th-232. In addition,
autoradiography of the soil particles in several size classes showed
the radioactivity in these samples was principally confined to
discrete particles within the soil matrix and not homogeneously
distributed on all particles.
SOLUBLE PHASE EQUILIBRIA
Dissolution of radioisotopes from the soil was examined with
laboratory leaching methods and quantified by means of a partitioning
coefficient, Kd, defined as:
KD =
mass of isotope/Kg solid
mass of isotope/Kg sea water
Four aliquots of the Middlesex soil sample B1A were suspended in
a known volume of 0.4 micron filtered lower Narragansett Bay sea
water. Samples were maintained at 20 degrees centigrade and
intermittently mixed over 20 hours. Particulate and soluble fractions
were then separated by filtration-(0.45 micron) and each fraction
analyzed for Ra-226, Pb-210, U-234, U-235, U-238, Th-227, Th-228,
Th-230, and Th-232 at EERF.
SETTLING VELOCITY DISTRIBUTIONS
Soil settling velocities were determined with a laboratory bench
scale settling column (Figure 2). The column was constructed from a 1
m long piece of 20 cm diameter PVC pipe. Transmissometers, interfaced
to a microcomputer, were mounted externally on the side of the column
(with glass windows) at depths of 16, 39, 62, and 85 cm to measure the
attenuation of infrared light as a function of time caused by the
-------
-20 cm •
SHUTTER ASSEMBLY
TRANSMISSOMETER
SAMPLING PORT
DRAIN PIPE
HIAC SENSOR
Figure 2. Laboratory settling column.
-------
passage of particles.
A HIAC-SSTA electronic particle sensor was mounted vertically on
the bottom plate of the column to determine the hydrated particle size
distributions. Soil samples were added instantaneously across the
entire surface area of the column by means of a gridded shutter, which
was located 3 mm above the water surface. The shutter was interfaced
with the microcomputer and triggered the transmissometer sampling when
opening. Filtered (0.4 micron) lower Narragansett Bay seawater was
added to the column before each experiment. The intensity of;each
transmissometer was matched by equalizing electronic outputs using
filtered sea water in the column.
It was assumed that the light attenuation was proportional to the
particle concentrations. This was tested by addition of four soil
samples of known mass to the column. The integrated area of the
transmissometer response versus time was found proportional to the
mass of material added. It was also assumed that the ratio of
attenuation to particle concentration was the same for all of the
particle sizes.
As the particles settled past a transmissometer, the attentuation
of the light beam increased as a direct function of the particle
concentration. The analog signal was passed through an
analog-to-digital converter, and the corresponding digital values were
stored. The response, i.e., particle concentration, at each depth was
integrated over time and normalized. A probability density function
was calculated to provide estimates of the percent of the total mass
passing a given depth as a function of time. The cumulative velocity
distribution, for a given depth was calculated from the empirical
relationship between cumulative sum of percent mass as a function of
depth divided by time. The vertical velocity distribution was,
therefore, determined by measuring the time of travel for each
fraction of total mass through a known distance (velocity = distance/time),
Probability and cumulative frequency curves were determined fbr each
experimental run.
MESOCOSM-SCALE EXPERIMENTS
A series of experiments in MERL mesocosms were designed to
quantify the effects of vertical dispersion on settling rates to
provide more realistic mass transport estimates for the model
predictions. These experiments were accomplished by modifying the
normal up/down plunger mixing used by MERL (Nixon et al., 1980) with
axially- driven impellers which rotated 360 degrees in a
clockwise/counterclockwise mode (Figure 3). This design provided
uniform mixing throughout the water column, as quantified by dye
studies.
The vertical dispersion in the mesocosm tanks was determined at
three levels of mixing. This was accomplished by first allowing the
impeller to operate at a set rate sufficiently long for a dynamic
equilibrium to be established in the tank. Rhodamine-WT dye was then
-------
B
Figure 3. MERL mesocosm. A. Normal up/down plunger
mixing. B. Axially driven impeller mixing.
a. Inflow to mesocosm. b. Plunger.
c. Sediments, d. Sampling ports for
suspended solids measurements. e. Axially
driven impeller. f. Sampling ports for
dye measurements, g. Impeller.
-------
sprayed onto the water surface and the concentration determined at 5
depths (0.5, 1.5, 2.5, 3.5, 4.5 m ) as a function of time. Dye
studies at mixing speeds of 0.37, 0.52, and 0.85 rad/sec gave
estimated dispersion coefficients of 8, 11, and 26 cm2/sec,
respectively. These data were used to calibrate a dispersion,model
(discussed later). Model predicted dye concentrations agree well with
measured concentrations (see, for example,, Figure 12).
.The effects of vertical dispersion on the transport of the soil
were tested by adding a known mass (ca. 4 kg) of soil, uniformly and
instantaneously, to the surface of the water column for a fixed
dispersion. Soil was added to the tank with a louvered aluminum fan
shutter (1.3 by 1.3 m) suspended 10 cm above the water surface.
Particle mass was intermittently sampled from three 2.5 cm diameter
syphon tubes, 1.2, 2.2 and 4.4 m below the water surface. The
hydrostatic head of the water column was sufficient to provide flow
rates of 0.9 I/sec. Samples (4.0 1) were taken on a log time scale
for up to 4 hours after a soil addition.
Particulates were separated from solution by filtration.
Particles greater than 63 microns were separated with a sieve, rinsed
onto a filter, and total mass for the known volume calculated. The
mass of less than 63 micron particles was determined by collecting a
500 ml aliquot of water after passing the sieve, then quantifying the
mass retained on passage through a 1 micron filter. The mass per unit
volume was calculated and added to the greater than 63 micron fraction
to give the concentration of total particulates. Water and
particulate samples were collected for radioisotope analysis from the
11 cmVsec dispersion experiment. The percentage of mass removal
from the water column was calculated by volume weighting the total
suspended solids concentration for each sample period, summing for the
sample period, and dividing by the known initial mass.
10
-------
SECTION 3
EXPERIMENTAL RESULTS
SIZE DISTRIBUTION
The grain size of the Middlsex soil ranged from less than 63
microns to greater than 2000 microns. Seventy-five per cent of the
mass was measured between 63 and 2000 microns. Cumulative mass
fraction versus particle size showed that this sample of Middlesex
soil was dominated by sand size particles smaller than 1 mm (Figure
4). The median grain size was approximately 350 microns.
Comparison of bulk sample size classifications between 20 and 600
microns (this range represents 62 per cent of the total mass) by the
wet and dry methods indicated that apparently larger sizes were
measured with the wet sizing technique (Figure 5). This may have
resulted from mechanisms such as particle agglomeration, surface
wetting, and swelling from hydration, as well as from differences in
the measuring techniques.
Comparision of dry and wet size distribution data for particles
from single dry sieving size classes (63-90, 90-106, 106-180, and
180-355 microns) supported the apparent size shifts (Figures 6a-e).
This tendancy was less pronounced for the larger size classes than for
the smaller size classes.
ACTIVITY DISTRIBUTIONS
Gross estimates for radioactivity (total gamma plus beta or total
alpha) indicated significant activity was associated with each size
class (Figure 7), but the activity was not uniformly distributed among
size classes on a unit mass basis (Table 1). Furthermore,
autoradiography of the soils demonstrated the activity was associated
with discrete particles. The median in gross activity was found at
the 125 micron size Figure 7).
The mass-weighted cumulative activity distributions for
individual isotopes were different from the cumulative mass and the
cumulative gross activity distributions (Figure 8). The individual
isotopes that were measured exhibited higher median sizes for activity
than did the measure of gross activity. Ra-226, U-234, U-238, and
Th-230 were similar to each other, with median distributions at 250
microns. Pb-210 and Po-210 had higher median distributions, 300 and
600 microns, respectively. Several isotopes, Th-227, Th-228, Th-232,
and U-235, did not show consistent cumulative activity distributions
due to large counting errors resulting from low activity and small
sample sizes.
Less than 10 per cent of the activity was found associated with
the silt-clay fraction (less than 63 microns) of the sample (Figure
8). Independent of the method of quantification and isotope of
interest, the radioactivity associated with the sample resided in the
11
-------
N>
c
u
n
u
L
A
T
I
V
E
F
R
A
C
T
I
0
N
0*8 .
0*6 .
0.4 -
0.2 ,
0
0
250 500 750 1000 1250
SIZE (MICRONS)
1 T
1500 1750 2000
Figure 4. Cumulative mass versus particle size.
-------
c
u
n
u
L
A
T
I
v
E
n
A
s
s
F
R
A
C
T
I
0
0,8 _
0.6 .
0.4 _
0.8 .
0
20-40 40-100 100-200 200-300
PARTICLE SIZE (MICRONS)
I
300-600
Figure 5. Size distribution between 20 and 600 microns as determined by
wet and dry methods.
-------
n
A
s
s
F
R
A
C
T
I
0
N
wet
dry
0.8 .
0*6 _
0*4 -
0*8 .
0
T
T
20-40 40-100 100-200 800-300 300-800
PARTICLE SIZE (MICRONS)
Figure 6. Size distributions of particles separated into size classes by
dry sieving, as determined by wet and dry methods. (a) 63-90
micron size class.
-------
n
A
s
s
F
R
A
C
T
I
0
N
wet
drg
0.8 -
0.6 -
0.4 _
0.2 .
0
20-40 40-100 100-200 200-300
PARTICLE SIZE (MICRONS)
300-600
Figure 6b. 90-106 micron size class.
-------
N
A
S
S
F
R
A
C
T
I
0
N
wet
drg
0*8 .
0,6 _
0.4 -
0.2 -
0
- , - \
80-40 40-100 100-300 200-300
PARTICLE SIZE (MICRONS)
300-600
Figure 6c. 106-250 micron size class.
-------
n
A
S
S
F
R
A
C
T
I
0
N
wet
dry
0.8 _
0.6 .
0.4 .
0
20-40 40-100 100-200 200-300
PARTICLE SIZE (MICRONS)
300-600
Figure 6d. 250-355 micron size class.
-------
H
00
M
A
S
S
F
R
A
C
T
I
0
N
wet
dry
0*8 .
0*6 .
0*4 -
0*2 .
0
I I I
20-40 40-100 100-200 200-300
PARTICLE SIZE (MICRONS)
300-600
Figure 6e. 355-600 micron size class.
-------
c
u
n
u
L
A
T
I
u
E
F
R
A
C
T
I
0
0*8 -
0.4 _
0.2 _,
0
0
E
HASS
GROSS ACTIVITY
I
250
500
750 1000 1250
SIZE (MICRONS)
1500 1750 2000
Figure 7. Cumulative gross activity versus particle size.
-------
TABLE 1. RADIOISOTOPE ACTIVITY FOR MIDDLESEX SOIL**
Particle
Diameter
<63
63
75
90
106
125
180
250
355
500
710
1000
>2000
Total
• Ra-226
:+ 2-3*
50.3
8.6
14.2
15.1
13.6
59.3
28.5
24.9
20.8
18.8
17.3
58.3
29.1
358.8
Pb-210
14-41*
56.0
8.4
19.1
16.4
14.9
61.5
27.7
29.9
19.9
29.2
32.5
149.3
49.6
514.5
Po-210
7-12*
57.2
8.4
13.4
13.7
15.9
52.9
37.7
27.2
19.3
20.0
26.3
64.7
31.8
388.3
U-234
12-16*
49.2
9.7
11.4
17.3
14.8
65.4
26.6
30.1
14.2
24.8
35.1
53.7
24.1
376.5
U-235
17-49*
4.6
3.8
2.3
0.8
1.8
2.4
1.1
1.8
0.7
1.1
1.9
2.2
0.9
25.5
U-238
12-16*
49.1
9.7
11.8
16.2
14.0
63.9
28.1
28.5
15.2
24.7
33.5
56.2
22.2
373.2
Th-227
26-62*
4.3
0.7
0.9
1.0
1.1
3.9
2.3
1.7
1.2
2.2
22.1
9.7
3.2
54.2
Th-228
10-218''
7.0
1.1
1.3
1.6
1.5
4.2
1.9
0.9
0.5
<0.1
0.3
52.9
<0.1
73.1
,
Th-230
'< 6-15*
• 54.9
, 10.2
11.2
15.2
13.5
57.9
31.6
21.7
16.6
23.5
31.6
84.4
31.9
404.1
Th-232
8-69*
8.2
1.2
1.5
1.3
1.7
47.9
2.0
1.3
1.3
1.3
0.3
76.0
1.3
145.2
4- Particle diameter in microns. Particles separated by standard sieves.
* Per cent counting error specified as two standard deviations.
** Activity express in pCi/g times mass fraction in that sieve size (g/g).
20
-------
l-o
c
u
I
0*8
0*6
0*4 .
0
RA-aae
a—0PB-310
0
850 500
750 1000 1250
SIZE (MICRONS)
1500 1750 a000
Figure 8. Cumulative individual isotope activity versus particle size.
-------
larger sand size particles. Therefore, with the exception of the
fraction of the isotopes mobilized to the dissolved phase, transport
of the isotopes should be described by processes controlling the
transport of sand size and larger particles.
ISOTOPE DISSOLUTION
Partition coefficients (Kd) for the individual isotopes ranged
between 10^ and 10^ (Table 2), depending on the isotope. The
isotopes which undergo the most mobilization to the dissolved phase
are Ra-226 and the uranium isotopes. Based on these experiments, 20
to 24 percent of these isotopes may be converted to the soluble phase
within 20 hours of initial contact with sea water. Considering the
small sample size, sandy texture, association of activity with
discrete particles in these samples, and counting errors, it seems
likely that these estimates may not accurately represent the
dissolution which may have been expected from these soils. Mass
balance considerations from the mesocosm transport studies suggested
that 10 percent of the uranium and Ra-226 may be released to the
dissolved phase while less than 1 percent of the Po, Th and Pb will be
solubilized over periods of hours to days (Table 3). Dissolution over
time scales up to 3 months are being examined using mesocosm studies
(Hunt, 1986). These results suggest that most of the radioactivity
(90 percent) will remain associated with the soils as they are
transported through the water column.
SETTLING VELOCITY DISTRIBUTIONS
Soil settling velocities were determined for a series of undried
composite Middlesex samples and a series of selected size classes
which had been dried. Reproducibility between runs of a given sample
was good. An example of the results from a typical experimental run
is shown in Figure 9. (See Appendix A for data from all of the
experimental runs.) A significant fraction (20 per cent) of the
composite sample settled at greater than 5 cm/sec in the upper 16 cm
of the column. However, the fraction of particles descending past 39,
62, and 85 cm at velocities greater than 5 cm/sec decreased to 10 per
cent. The momentum from the actual soil addition could have resulted
in the observed decrease. Other causes may be theorized for this
observed shift, such as wetting or hydration of particles and
disaggregation of soil clumps. Regardless of the cause, the shift
to a slower velocity reached an equilibrium at 39 cm.
Minimum settling velocities for the composite sample increased
from 0.2 cm/sec at 16 cm to 1.3 cm/sec at 85 cm. The median settling
velocity followed a similar trend. The observed velocity distribution
shift suggested by these data had not reached an equilibrium 'state by
85 cm. Possible .causes for this observed shift include wetting or
hydration of particles and agglomeration of soil particles.
Settling velocity distributions for individual size classes also
increased on passage through the sea water (Figure lOa-e),
particularly in the smaller size classes. The changes in settling
22
-------
TABLE 2. ISOTOPE SOLUBILITY AND PARTITION COEFFICIENTS FOR MIDDLESEX SOIL
NJ
Suspended
solids
KD+
270
per
mg/kg
cent
error*
soluble
Ra-226
Pb-210
Po-210
U-234
U-235
U-238
Th-227
Th-228
Th-230
Th-232
8.98
1.37
3.99
9.59
1.16
9.11
9.18
5.78
1.26
1.19
3
10
5
10
5
10
3
10
4
10
3
10
4
10
4
10
6
10
6
10
28
2.5
0.9
27
24
28
3.7
5.8
0.3
0.3
1
70
29
9.5
19
9.5
62
34
35
140
280
KD+ per
mg/kg
cent
error*
900
KD+ per
soluble
3
4.43 10
5
1.38 10
5
2.08 10
4
1.33 10
4
1.10 10
4
1.42 10
5
1.20 10
4
4.46 10
5
8.50 10
6
1.63 10
44
2.5
1.6
21
24
20
2.8
7.2
0.4
0.2
1
91
20
13
21
13
77
32
33
140
3.84
7.11
5.09
4.34
4.34
4.53
7.41
2.83
9.91
2.86
mg/kg
cent
error*
KD+
soluble
3
10
4
10
5
10
3
10
3
10
3
10
4
10
4
10
5
10
7
10
22
1.5
0.2
19
20
19
1.5
3.7
0.1
<0.1
.7
70
30
13
15
13
65
26
36
2500
4.37
1.82
2.77
4.95
3.44
4.74
5.44
3.82
9.55
2.85
1220 mg/kg
per
cent
error*
soluble
4
10
5
10
5
10
3
10
3
10
3
10
4
10
4
10
5
10
6
10
2
0.4
0.2
14
19
14
1.5
2.0
<0.1
<0.1
2
119
21
14
16
14
51
31
34
300
+ Partition coefficient (kg seawater/kg particulate).
* Per cent counting error specified as two standard deviations.
-------
TABLE 3. MASS BALANCE ESTIMATES OF ISOTOPE SOLUBILITY FOR MIDDLESEX
SOIL FROM MERL MESOCOSM EXPERIMENTS
ISOTOPE
Ra-226
Pb-210
Po-210
U-234
U-235
U-238
Th-227
Th-230
Th-232
TOTAL
ADDED
(pCi)
1676
1734
1593
1584 ,
44
1562
119
1804
1848
TOTAL
SOLUBILIZED*
(pCi)
120
18
21
135
6
132
0.62
1.7
1.3
PERCENT
SOLUBILIZED
7.2
1.0
1.3
8.5
13.6
8.5
0.5
0.09
0.07
* After 4 hours of exposure.
24
-------
N)
Ul
c
u
n
u
L
A
T
I
v
E
F
R
A
C
T
I
0
N
0.8 -
0.6 _
0.4 _
0
0
Depths in Settling Column
. 16 cm
39 cm
—•- — 62 cm
85 cm
I
2
l
3
l
4
i
5
l
6
l
8
l
9
10
UELOCITY, H/T (CM/SEC)
Figure 9. Settling velocity distribution for Middlesex soil determined from
laboratory column experiments.
-------
N3
1 -.
c
u
n
u
L *
A
T
I
U
E 0*6 .
n
A
S
8 0.4.
F
R
A
J 0*2 .
0
N
•
<
*
/ /
/ /
/ /
' '
I /
/
J i
1 ''
•7 !
1 1
i
/i
M
/ Depth in Settling Column
1 16 cm
j 39 cm
• 62 cm
j 85 cm
I
1 1 1 1 1 1 1 1 i
123456789 1<
UELOCITV, H/T (Cn/SEC)
Figure 10. Settling velocity distributions for Middlesex soil separated into
size classes by dry sieving, (a) 63-90 micron size class.
-------
Ni
c
u
n
u
L
A
T
I
u
E
n
A
s
s
F
R
A
C
T
I
0
N
0.8 .
0.6
0*4 _
0.3 _
Depth in Settling Column
16 cm
39 cm
62 cm
85 cm
0
0
T
3
"T
4
T
5
T
6
"T
7
UELOCITY, H/T (CM/SEC)
8
10
Figure lOb. 90-106 micron size class.
-------
S3
OO
c
u
n
u
L
A
T
I
y
E
n
A
s
s
F
R
A
C
T
I
0
N
0,8 .
0*6 -
0.4 _
0.3 .
0
T
1
T
S
Depth in Settling Column
16 cm
39 cm
62 cm
85 cm
T
3
"T
4
T
5
T
7
8
10
VELOCITY, H/T (CI1/SEC)
Figure lOc. 106-250 micron size class.
-------
N3
c
u
n
u
L
A
T
I
u
E
n
A
s
s
F
R
A
C
T
I
0
N
0*8 .
0*6 .
0.4 .
0*2 .
0
0
Depth in Settling Column
39 cm
62 cm
85 cm
I 1 1
45 6
VELOCITY, H/T (CM/SEC)
T
7
8
"T
9
10
Figure lOd. 250-355 micron size class.
-------
00
o
c
u
n
u
L
A
T
I
U
E
n
A
s
5
F
R
A
C
T
I
0
N
0*8 -
0.4 _
0.2 .
0
0
T
i
cm
85 cm
T
T
T
T
3 4 S 6 7
VELOCITY, H/T (CM/SEC)
8
10
Figure lOe. 355-600 micron size class.
-------
velocity distributions at the four transmissometer depths indicated
that an equilibrium state had not been reached by 85 cm for the three
smaller size classes.
The settling column results indicated that the Middlesex soil
will fall rapidly through the water column, and that changes in
settling velocity distribution would occur upon introduction into sea
water. Whatever the mechanism for the observed change, the increased
velocities act to reduce the water column residence time for the finer
particles. The velocity distributions approached an equilibrium state
within very short vertical distances, particularly relative to the
depth at a deep-ocean disposal site.
MESOCOSM-SCALE EXPERIMENTS
The effects of vertical dispersion on the transport of the soil
in the mesocosm tank was difficult to determine. The rapid descent of
the particles in the MERL mesocosm limited the quality of data.
Approximately 4 seconds was required per sampling to obtain the 4.0
liters used to determine the solids concentration. A peak in mass
descent through the water column was not detected during the
experiments (see Figure 11). The peak in mass was most probably
between ports for the initial samplings. We were able to estimate
the total mass of particles remaining in the mesocosra as a function
of time to determine the mass removal from the water during the
experiments. These results (Table 4) suggest changes in dispersion
influenced the mass removal in the 5 m MERL water column, with
higher dispersion increasing the mass residence time. Differences
in mass removal at different dispersions could be accounted for by an
analytical expression relating gravitational settling, V, and vertical
diffusion, D (Table 4). Extrapolating this analytical expression to
a 4000 meter water column suggests that the dispersion levels, as
used in the MERL tank, would modify mass removal times by less than
five per cent.
31
-------
N3
c
0
N
C
E
N
T
R
A
T
I
0
N
P
P
T
-0.5.
10
-1.5.
10
-a.5.
-3.
10
10~3*5
.-4
10
10
0
10
118 CM
—a 230 cn
CM
10
TII1E (SECONDS)
10
Figure 11. Mass removal experimental results for particles greater than 63
microns and with 11 cm /sec dispersion in MERL mesocosm.
-------
TABLE 4. EFFECTS OF DISPERSION ON MASS REMOVAL OF MIDDLESEX SOIL
FROM MERL MESOCOSM EXPERIMENTS
Dispersion
— •— —-— — — — ~— —^-—— _ ._ .. . _ r—lr.-M-ra-iMi-m -!•••_•—•--•—»«» .^ i, i, • . • I I
2 2
11 cm /sec mixing 26 cm /sec mixing
per cent • ~
mass removed t+ (sec) V* (cm/sec) t+ (sec) V* (cm/sec)
50 94 5.3 78 6.4
84 229 2.5 363 1.7
90 385 1.6 611 1.2
95 1322 .59 2543 .43
+• Estimated by volume weighting concentrations for a sampling period,
and linearly extrapolating between sampling periods.
* V is estimated from mass removal time by the following expression
(derived from results in Csanady, 1973):
X = 0.5 [erf((b-Vt)/(2 / Dt)) + erf((b+Vt)/(2 /~Dt»],
where
X = fraction remaining in water column,
erf( ) = error function,
b - depth of water column,
V = settling velocity,
t = time for removal, and
D = vertical diffusion.
33
-------
SECTION 4
MODELING APPROACH
MODEL DEVELOPMENT
Particulate matter is vertically transported through the water
column by a combination of convection, settling, and diffusion (both
turbulent and molecular) processes. The model used here assumes that
vertical convection of the fluid was not a significant transport
process. It has been shown that vertical convection is only
significant in large water bodies where strong upwelling and
downwelling occur, e.g., in nearshore coastal regions, while settling
and diffusion are ubiquitous phenomena (Lick, 1982). Two different
levels of models were considered to describe vertical transport of
non-interacting particles. Level 1 assumed the particles can be
characterized by a single average particle size and an associated
average settling velocity. The Level 1 particles will be referred to
as monodispersed. Level 2 assumed the particles can be treated as a
system of independent, or non-interacting, size classes, each with its
own characteristic settling velocity. Level 2 particles will be
referred to as polydispersed.
Level 1
The one-dimensional, time-dependent, convective-diffusive
transport model presented here involved two vertical transport terms,
settling and diffusion. The settling term for non-interacting,
single-sized particles was the mean settling velocity of the material
relative to the fluid, V (the velocity of the fluid, Vf, minus the
particle fall velocity, Vp), multiplied by the concentration:
3C
at
avc
(1)
where C is the particle concentration, z is depth, and t is time. The
particle fall velocity, Vp, was expressed as a function of particle
diameter for three different Reynolds number regimes (Weber, 1972):
Stokes settling (Re < 1):
_g 2
18y F
Vp
(2)
34
-------
Transitional settling (1 < Re < 1000):
1.6
0.72 (pp—po) D
Vp , ;—-—7 ,
(3)
Newton's settling (1000 < Re < 25000);
Vp =» 1.82
(pp-po) D S
(4)
Po
where Re is the Reynolds number based upon particle diameter (DpoVp/y),
g is gravitational acceleration, y is viscosity of the fluid, pp is
particle density, Po is fluid density, and D is particle diameter.
Alternatively, Vp was determined from laboratory settling column
experiments.
The vertical diffusion was a combination of the vertical eddy
diffusivity coefficient, Dv, and the molecular diffusion coefficient,
Dm,
3C
• •••••
3t
a ' 3C
(Dv + Dm)
3z 3z
(5)
The equation for the first level vertical transport model was the
combination of Equations (1) and (5),
3G 3
•--- = -- (Dv + Dm)
3t 92
3C
3VC
3z
(6)
The boundary conditions used with Equation (6) were that (1) the upper
boundary was treated as a surface with no net mass flux across it, and
(2) the bottom boundary was treated as an adsorptive boundary, i.e.,
particles that reach this boundary "stick" to it. The initial
condition was specified as having the entire mass of soil particles in
the surface water layer, i.e. the particles were considered to be
added instantaneously to the surface at time t = 0.
There are several assumptions that govern our use of Equation
(6). These are:
1. The horizontal variation in concentration was negligible.
35
-------
2. Mixing was accounted for by a constant eddy diffusivity.
3. There was no hindered or non-discrete settling of particles
due to hydrodynamic influences.
4. Vertical convective fluid transport was insignificant, i.e.,
Vf - 0.
5. The particle size distribution was treated as a single-size
category characterized by an average particle size with an
average settling velocity.
6. The particles formed a stable suspension, i.e., no
coagulation occurred.
Level 2
Most particulate systems, including the Middlesex soil, are not
monodispersed suspensions. Under the condition where assumption (5)
above is not valid, a system of equations must be employed, each
similar to Equation (6). Each equation would correspond to a single
size category. The Level 2 modeling approach assumed that several
state variables, one for each representative size or velocity
category, could be used to describe the vertical convective-diffusive
transport of the polydispersed suspension.
In addition to the assumptions for the Level 1 model, there was
an additional assumption for the use of the Level 2
convective-diffusive model. This assumption pertained to the dynamic
behavior of the size distribution. It was assumed that the particles
neither coagulate and form larger particles nor break-up and form
smaller particles. This assumption is only valid for stable,
non-coagulating systems that are approaching or have reached a
particle size distribution equilibrium, i.e., the particle formation
and break-up balance.
A system of partial differential equations was used to describe
the vertical transport of the polydispersed particles. This system
is:
3Ci 3 3Ci 3ViCi
= (Dvi + Dmi) -
3t 3z 3z 3z
1=1, ... N, (7)
where Ci is the particle concentration in a size category i; Vi, Dvi,
and Dmi are the mean settling velocity, vertical eddy diffusivity
coefficient, and molecular diffusivity coefficient, respectively, for
size category i; and N is the number of size categories. Boundary
conditions and the initial conditions were similar to those used for
the Level 1 model. The equations within the above system are not
coupled as they would be for a coagulating suspension of particles,
36
-------
i.e., for a dynamic particle size distrbiution. Therefore, the
equation for each size category was solved independently.
MODEL CALIBRATION AND DATA SYNTHESIS
The calibration of the dispersion model, Equation (5), was
conducted to quantify the vertical dispersion in a MERL tank. This
calibration was based on the data obtained during the dye studies.
Figure 12 shows an example of predicted model output superimposed onto
a plot of the observed data. The model output, with vertical
dispersion at 11 cm^/sec, agrees well with observed data from the
dye study conducted at a mixing speed of 0.52 rad/sec. Similar
agreement was obtained for each mixing level in the MERL tank. The
results of the other experiments appear in Appendix B.
Because the results of the laboratory bench scale experiments
indicated that the soil settling could be treated not as a single
settling velocity but as a distribution of settling velocities, we
applied the Level 2 modeling approach to synthesize the data and to
make model projections for the deep-ocean environment. The convection
terms for the Level 2 system of equations were determined empirically
from the data obtained from the settling column experiments. These
data represent settling velocity distributions for the polydispersed
suspension of soil particles.
The velocity distributions actually used in the model were those
measured at the greatest depth in the experimental settling column (85
cm). Apparent equilibrium was approached at this depth for the
settling velocities in the laboratory system. The equilibrium size
distribution assumption is, therefore, quite reasonable for the
projections to the deep-ocean environment. The size distribution of
the settling particles in the deep-ocean environment would reach an
equilibrium condition before descending through a significant fraction
of the total depth (4000 m). This equilibrium would occur, if for no
other reason, due to reduced particle-particle interaction caused by
simple particle number dilution as a result of the turbulent
dispersion and differential gravitational settling.
MODEL APPLICATION
The Level 2 convective-diffusive model was applied assuming
conditions that would exist at a hypothetical deep-ocean disposal site
in 4000 m of water. The model was used to estimate the soil particle
residence time in the water column and the distance along the
direction of mean flow to deposition on the bottom sediments. An
empirically determined settling velocity distribution and assumed
vertical dispersion, ranging from 0 to 100 cm^/sec, were used.
Figure 13 shows time of arrival at the bottom sediments for the
Middlesex soil assuming vertical dispersions of 0, 50, and 100
cm2/sec. The model predicts that at least 95 per cent of the soil
mass would arrive at the sediment surface in less than five days.
Figure 13 indicates that variations in vertical dispersion over the
range characteristic of deep-ocean sites has negligible effect on the
37
-------
CX5
X
H
OL
UJ
Q
TIME (min)
D
O
0.5
I I I T I I
1.5 2 2.5 3 3.5 4
RELATIVE CONCENTRATION
4.5
Figure 12. Dye experimental results with 11 cm2/sec dispersion in MERL mesocosm.
-------
c
u
n
u
L
A
T
I
M
A
S
S
F
R
A
C
T
I
0
N
0*8 .
0*6
0.4 _
0.3 .
0
0
0*5
T
Vertical Dispersion
0 cm /sec
50 cm2/sec
.100 cm2/sec
T
T
a
T
3
T
"T
4
1 1.5 3 8.5 3 3.5
TIME TO REACH BOTTOM SEDIMENTS (DAYS)
4.5
Figure 13. Time of arrival at the bottom of 4000 meter water column for
Middlesex soil.
-------
time of arrival for the Middlesex soil.
Information on the horizontal current structure off the northeast
coast of the United States (Ingham et al., 1977) was used to calculate
the horizontal distance particles would travel before they reach the
bottom. The downstream bottom impact distance from the disposal site
is shown in Figure 14 for assumed horizontal current velocities of
2.5, 5.0, 7.5 and 10.0 cm/sec. This figure indicates that the ;
distance these particles would travel, for the assumed site condition,
ranges from 0.5 to 40.0 km. O'Connor et al. (1985) reported a ;
vertically-weighted mean velocity based on the deep-ocean current
structure off the northeast coast of the United States. They
determined a mean velocity of approximately 2.0 cm/sec. With this
velocity, the impact distance along the direction of mean flow for 95
per cent of the mass would be aproximately 6 km. For practical
purposes, the disposed soil would deposit within the confines of a
typical disposal site based on a horizontal current of 2.0 cm/sec.
Our projections for a deep-ocean disposal site were based upon
experimental data for the Middlesex soil that had been excavated,
sieved, and homogenized. If an actual disposal operation were to
commence, the physical characteristics of the soil disposed would be
expected to be different than that of the homogenized soil we used.
However, our results should provide upper-bound estimates on the .time
of arrival at the bottom sediments and on the bottom impact distance
along the direction of mean flow.
40
-------
c
u
n
u
L
A
T
I
U
E
n
A
s
s
F
R
A
C
T
I
0
N
0.8 .
0.6 .
0.4 _
0.E _
0
Horizontal Current
2.5 cm/sec
..... 5.0 cm/sec
7.5 cm/sec
— — — 10.0 cm/sec
T
0
10 15 20 25 30 35
DISTANCE FROM DUUPSITE (KM)
45 50
Figure 14. Bottom impact distance along mean flow direction for Middlesex
soil disposed in 4000 meters of water.
-------
SECTION 5
CONCLUSIONS
A combined experimental and modeling approach was conducted
within the framework of the hazard assessment methodology developed
at ERLN. Experimentation was conducted to characterize Middlesex
soil in terms of particle size distribution, specific gravity,
radioactivity, and soluble phase equilibria. In addition, dynamic
settling velocity distributions were measured. The data derived
from the experimental work were synthesized using a modeling approach
that emphasized the vertical convective-diffusive transport of a
polydispersed suspension of particles. Ecosystem level studies with
Middlesex soil in mesocosms are presented in Hunt (1986). The
following conclusions may be drawn from our study:
1. The Middlesex soil in this study had 75 per cent of the mass
on particles between 63 and 2000 microns, had a median size
of 350 microns, and a bulk specific gravity of 2.31.
2. A measure of gross radioactivity(found activity associated
with particles sizes from less than 63 to greater than 2000
microns, with a median activity at 125 microns.
3. The specific element or isotope activity distributions showed
peaks in activity between 250 and 600 microns. The particle
size for median activity distribution of individual isotopes
was not coincident with either the size or gross activity
distributions.
4. The fraction of total activity which might solubilize upon
initial discharge into marine waters was at most 20-24 per
cent for the isotopes Ra-226, U-234, U-235, and U-238, and
less than 5 per cent for Pb-210, Po-210, and the other
isotopes.
5. The velocity distributions were initially dynamic, but tended
to converge with increasing depth in a one meter settling
column. This indicated that the soil suspension was shifting
to a more stable state, and that the velocity distribution
was approaching an equilibrium.
6. The measured soil settling velocity ranged up to 8.2 cm/sec,
with a median of 2.1 cm/sec.
7. Gravitational settling would be the dominant vertical
transport mechanism from Middlesex soil at a deep-ocean
disposal site, with vertical dispersion having less than a
5 per cent effect.
8. The Middlesex soil would impact the bottom sediment in 4000
meters of water at distances from 0.5 to 40 km along the
direction of mean flow, for horizontal velocities in the
42
-------
range of 2.5 to 10 cm/sec.
Ninety-five per cent of the Middlesex soil disposed in 4000 m
of water would arrive at the bottom in less than five days.
The particles with the largest settling velocities may reach
the bottom in approximately twelve hours.
43
-------
REFERENCES
Carver, R.E. (ed.) 1971. Procedures in sedimentary petrology. John
Wiley and Sons, New York.
Csanady, G.T. 1973. Turbulent diffusion in the environment.
D. Reidel Publishing Co., Boston.
Hunt, C.D. 1986. Fate and bioaccumulation of soil-associated
low-level natural radioactivity discharged into a marine
ecosystem. Report in preparation.
Kupferman, S.L., D.R. Anderson, L.H. Brush, L.S. Gomez, and
L.E. Shephard. 1984. FODOCS Annual Report March 1 - September
30, 1981. Rept. SAND82-0292, Sandia National Laboratories,
Albuquerque, New Mexico. 68 pp.
Ingham, M.C., J.J. Bisagni, and D. Mizenko. 1977. The general
physical oceanography of Deepwater Dumpsite 106. In: Baseline
Report of Environmental Conditions at Deepwater Dumpsite 106,
Vol. 1. NOAA Dumpsite Evaluation Rept. 77-1, U.S. Dept. of
Comm. Pub., pp. 29-54.
Lick, W. 1982. Entrainment, deposition, and transport of fine-grain
sediments in Lakes. Hydrobiologia, 91:31-40.
Nixon, S.W., D. Alonso, M.E.Q. Pilson, and B.A. Buckley. 1980.
Turbulent mixing in aquatic microcosms. In: Microcosms in
Ecological Research, G.P. Gisey (ed.). DOE Symposium Series,
Augusta, Georgia, Nov. 8-10, 1978. CONF-781101, NTIS, pp.
818-849.
O'Connor, T.P., H.A. Walker, J.F. Paul, and V.J. Bierman, Jr. 1985.
A strategy for monitoring of contaminant distributions resulting
from proposed sewage sludge disposal at the 106-Mile Ocean
Disposal Site. Marine Environmental Research, 16:127-150.
Prager, J.C., V.J. Bierman, Jr., J.F. Paul, and J.S. Bonner. 1984.
Hazard assessment of low level radioactive wastes. A proposed
approach to ocean dumping permit request analyses. Technical
Report, Environmental Research Laboratory, U.S. Environmental
Protection Agency, Narragansett, Rhode Island.
U.S. Congress P. L. 92-532. The Marine Protection, Research and
Sanctuaries Act of 1972, 33 U. S. 1401 et. seq.
Weber, W.J., Jr. 1972. Physiochemical processes for water quality
control. Wiley-Interscience, New York.
44
-------
APPENDIX A
BENCH-SCALE SETTLING VELOCITY DATA
45
-------
c
u
M
u
L
A
T
I
U
E
F
R
A
C
T
I
0
N
0.8 _
0.6 .
0.4 .
0.2 .
0
0
T
1
85 cm
T
T
34567
UELOCITY, H/T (CM/SEC)
8
T
9
10
Figure A-l. Settling velocity distribution for Middlesex soil determined from
laboratory column experiments.
-------
c
u
n
u
L
A
T
I
u
E
F
R
A
C
T
I
0
N
0,8 .
0.6 .
0.4 .
0.3 .
0
0
Depth in Settling Column
.1 16 cm
39 cm
62 cm
85 cm
I
a
r
A
i
5
I
6
I
7
r
8
i
9
10
VELOCITY, H/T (CH/SEC)
Figure A-2. Settling velocity distribution for Middlesex soil determined from
laboratory column experiments.
-------
oo
c
u
n
u
L
ft
T
I
u
E
F
R
A
C
T
I
0
N
0.8 .
0.6 -
0.4 .
0.S .
0
0
Depth in Settling Column
16 cm
39 cm
62 cm
85 cm
T
a
T
3
T
4
T
5
T
7
T
8
VELOCITY, H/T (CM/SEC)
T
9
10
Figure A-3. Settling velocity distribution for Middlesex soil determined from
laboratory column experiments.
-------
VD
c
u
n
u
L
A
T
I
u
E
F
R
A
C
T
I
0
N
0.6 .
0*4 .
0.8 .
0
0
Depth in Settling Column
16 cm
39 cm
62 cm
85 cm
"T
3
T
4
"T
5
T
6
~r
7
T
8
10
UELOCITY, H/T (CM/SEC)
Figure A-4. Settling velocity distribution for Middlesex soil determined from
laboratory column experiments.
-------
APPENDIX B
MESOCOSM-SCALE DATA
50
-------
E -
0.
UJ
Q
0 0.5
2.5 3 3.5
RELATIVE CONCENTRATION
Figure B-l. Dye experimental results with 8 cm2/sec dispersion in MERL mesocosm.
-------
0
Ui
N3
-I -
-2 -
x
H-
Q.
LU
Q
-4 -
-5
0 0.5
T
1.5 2 2.5 3 3.5
RELATIVE CONCENTRATION
Figure B-2. Dye experimental results with 11 cm2/sec dispersion in MERL mesocosm.
-------
Ul
OJ
E -
Q.
Id
Q
-5
0 0.5 I
!
2
!
3
2.5 3 3.5
RELATIVE CONCENTRATION
!
4
4.5
Figure B-3. Dye experimental results with 26 cm2/sec dispersion in MERL mesocosm.
-------
Ul
c
0
N
C
E
N
T
R
A
T
I
0
N
P
P
T
10
i
10
0.
10
-1-
10
-2.
10
-3J
10
-4
10
G-.
118 CM
—0280 CM
CH
10
—i \ r
102 102*5 103
TINE (SECONDS)
10
3,5
10
Figure B-4. Mass removal experimental results for particles greater than
63 microns and with 26 cm2/sec dispersion in MERL mesocosm.
-------
Ul
Ul
0
N
C
E
N
T
R
A
T
I
0
N
P
P
T
0
10
10
-0.5.
10
-1-
10
•1.5.
10
10
-a.5_
10
-3
10
118 cn
cn
cn
10
1.5
I I I
102 102*5 103
TIME (SECONDS)
10
3.5
10
Figure B-5. Mass removal experimental results for particles less than 63 microns
and with 26 cm /sec dispersion in MERL mesocosm.
-------
Ul
a\
c
0
N
C
E
N
T
R
ft
T
I
0
N
P
P
T
10
10*
0.5
100
-0.5.
10
-1.5.
10
-2.
10
-2.5.
10
-3
10
118 CH
220 CH
en
10
1.5
I 1 I
102 102*5 103
TIME (SECONDS)
10
3.5
10
4
Figure B-6. Mass removal experimental results for total mass and with 26 cm2/sec
dispersion in MERL mesocosm.
-------
Ui
c
0
N
C
E
N
T
R
A
T
I
0
N
P
P
T
10
10
10*
-0*5.
10-1-
-1.5
10
-a.5.
10
10
-3.5_
10
-4
10
0
10
•a
118 cn
en
cn
10
TIME (SECONDS)
10
Figure B-7. Mass removal experimental results for particles greater than
63 microns and with 11 cm2/sec dispersion in MERL mesocosm.
-------
Ui
oo
c
0
N
C
E
N
T
R
ft
T
I
0
N
P
P
T
0
10
10
-0*5.
10
-1.
10
-1*5.
10
-2.
10
-2*5.
10
-3
10
0
10J
118 CH
a—aa20 OH
«—M42 CM
T
.3
10
TIHE (SECONDS)
10
Figure B-8. Mass removal experimental results for particles less than 63
microns and with 11 cm2/sec dispersion in MERL mesocosm.
-------
m
vo
c
0
N
C
E
N
T
R
A
T
I
0
N
P
P
T
10
e
-0.5.
10
-i.
10
-1*5.
10
-a.
10
-2*5.
10
-3
10
0
10
118 en
—0220 en
en
—I
ie2
TIME (SECONDS)
3
10
r
4
10'
Figure B-9. Mass removal experimental results for total mass and with 11 cm2/sec
dispersion in MERL mesocosm.
-------
CTv
O
c
0
N
C
E
N
T
R
ft
T
I
0
N
P
P
T
10
-0*5.
10
-1.5.
3
10
10~2*5
-3
10
10
-3*5.
10
-4
10
0
10
118 cn
aaa0 cn
M43 CM
1
102
TIME (SECONDS)
10
Figure B-10. Mass removal experimental results for particles greater than
63 microns and with 8 cm2/sec dispersion in MERL mesocosm.
------- |