Unitecf States
Environmental Protection
Agency
Technology Transfer
EPA/625/4-89/019
[>EPA Seminar Publication
Transport and Fate of
Contaminants in the
Subsurface
-------
-------
Technology Transfer EPA/625/4-89/019 - Sept. 1989
Seminar Publication
Transport and
Fate of Contaminants in the
Subsurface
Center for Environmental Research Information
Cincinnati, OH 45268
and
Robert S. Kerr Environmental Research Laboratory
Ada, OK 74820
-------
NOTICE
The information in this document has been funded wholly or in part by the United States
Environmental Protection Agency under Contract 68-C8-0014 to Eastern Research Group, Inc. It
has been subject to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
11
-------
CONTENTS
Page
Chapter 1. Introduction [[[ 1
Chapter 2. Physical Processes Controlling the Transport of Contaminants
in the Aqueous Phase .............. . . . ........... ; . . .......... 5
Chapter 3. Physical Processes Controlling the Transport of Non-Aqueous
Phase Liquids in the Subsurface .................. ........... . 23
Chapter 4. Determination of Physical Transport Parameters ......................... 29
Chapter 5. Subsurface Chemical Processes .......................................... 41
Chapter 6. Subsurface Chemical Processes: Field Examples ........ ............. 57
Chapter 7. Microbial Ecology and Pollutant Biodegradation in Subsurface
Ecosystems [[[ gy
Chapter 8. Microbiological Principles Influencing the Biorestoration of Aquifers ........ 85
Chapter 9. Modeling Subsurface Contaminant Transport and Fate ............ . ..... 101
-------
ACKNOWLEDGEMENTS
This publication is based on a series of technology transfer seminars that were conducted in all
ten EPA Regions between October 1987 and February 1988. The seminars provided regulators
and technical specialists with a brief but intensive overview of the physical, chemical, arid
biological processes that govern the transport and fate of contaminants in the subsurface. A
secondary purpose of the seminar was to provide a summary of modeling approaches used to
make predictions about the transport and fate of contaminants in the subsurface and to describe
the current and potential regulatory uses of such models.
The EPA Center for Environmental Research Information (CERI) and Robert S. Kerr
Environmental Research Laboratory (RSKERL) developed this project to assist the technical
support and technology transfer efforts of the EPA Office of Solid Waste and Emergency
Response (OSWER) in the area of subsurface remediation. Seminar development and
coordination were aided by numerous personnel, representing the EPA Office of Research and
Development (ORD), OSWER, and the EPA Regions. Principal contributors to the project
include:
Authors:
Richard L. Johnson, Oregon Graduate Center, Beaverton, Oregon
Joseph F. Keely, Ground-water Quality Consultant, Portland, Oregon
Carl D. Palmer, Oregon Graduate Center, Beaverton, Oregon
Joseph M. Suflita, University of Oklahoma, Norman, Oklahoma
William Fish, Oregon Graduate Center, Beaverton, Oregon
Seminar Speakers:
Johnson, Keely, Palmer, and Suflita
Dermont Bouchard, ORD/RSKERL, Ada, Oklahoma
Michael Henson, ORD/RSKERL, Ada, Oklahoma
Technical Reviewers:
Jack Keeley, Consultant, Ada, Oklahoma
Ronald C. Sims, Utah State University, Logan, Utah
David Ostendorf, University of Massachusetts, Amherst, Massachusetts
Project Managers:
Carol Grove, ORD/CERI, Cincinnati, Ohio
Chuck Marshall, JACA Corp., Ft. Washington, Pennsylvania
Marion R. Scalf, ORD/RSKERL, Ada, Oklahoma
Mark Lennon and Sylvie Stanish, Eastern Research Group, Inc., Arlington, Massachusetts
Seminar Development/Coordination:
ORD/RSKERL: Don Draper, John Matthews
ORD/OEPER: Will LaVeille
OSWER: Janette Hansen, Victor Hays, Amy Mills, Charles Perry, Peter Tong, Burnell
Vincent
OGWP: Carey Carpenter
Region I: David Lang, Stephen Mangion, John Zannos
Region II: Ken Wenz
IV
-------
Region III: Eileen Burrows, Dom DiGivlio, Mindi Snoparshy, Gerallyn Vails
Region IV: Elmer Akin
Region V: Ken Chin, Dale Helmers, Ken Westlake
Region VI: Debbie Wright
Region VII: Jack Coakley
Region VIII: Charles Brinkman
Region IX: John Duff
-------
-------
CHAPTER 1
INTRODUCTION
Joseph F. Keely
Congress requires that the U.S. Environmental
Protection Agency (EPA), as well as other regulatory
entities and the regulated community, meet four
interrelated objectives for the protection of ground-
water quality. These objectives are:
1. Assessment of the probable impact of existing
pollution on ground water at points of
withdrawal or discharge (Safe Drinking Water
Act of 1974 & 1986 (SDWA)).
2. Establishment of criteria for location, design,
and operation of waste disposal activities to
prevent contamination of ground water, or
movement of contaminants to points of
withdrawal or discharge (Resource Conservation
and Recovery Act of 1976 (RCRA), and the
Hazardous and Solid Waste Amendments of 1984
(HSWA)).
3. Regulation of the production, use, and disposal of
specific chemicals possessing an unacceptably
high potential for contaminating ground water
when released to the environment (Toxic
Substances Control Act (TSCA), and the Federal
Insecticide, Fungicide, and Rodenticide Act
(FIFRA)).
4. Development of remediation technologies that
are effective in protecting and restoring ground
water quality without being unnecessarily
complex or costly and without unduly restricting
other land use activities (Comprehensive
Environmental Response, Compensation, and
Liability Act of 1980 (CERCLA or Superfund),
and the Superfund Amendments and Reauthori-
zation Act of 1986 (SARA)).
To achieve these objectives, definitive knowledge of
the transport and fate of contaminants in the
subsurface environment is essential. Without this
knowledge, regulatory agencies (such as EPA) run
the twin risks of under-control and over-control.
Regulatory under-control would result in inadequate
prevention and cleanup of ground-water contami-
nation. Regulatory over-control would result in
costly preventative actions and remedial responses
to contamination. However, gaining and using
knowledge about contaminant transport and fate can
be difficult because of the complexity of the sub-
surface environment. The activities of site character-
ization and remediation can be used to illustrate this
complexity.
Site Characterization
Transport and fate assessments require inter-
disciplinary analyses and interpretations because
the processes involved in these activities are
naturally intertwined. Examining each process in
isolation is much like taking photographs of an
object from different perspectives and then trying to
piece them together to describe the object. Each
transport process must be viewed from the broadest
of interdisciplinary viewpoints, and the interactions
between them identified and understood. In addition
to a sound conceptual basis, integrating information
on geologic, hydrologic, chemical, and biological
processes into an effective contaminant transport
evaluation requires data that are accurate, precise,
and appropriate at the intended problem scale.
The issues of contaminant transport and fate in the
subsurface are particularly difficult to address at
Superfund sites because of the complex array of the
chemical wastes involved. The hydrogeologic
settings of these sites are usually measured in
-------
hundreds of feet and, at this scale, are extremely
complicated when characterized for a remediation
plan. The methods and tools used for large-scale
characterizations are generally applicable to the
specialized needs at hazardous waste sites; however,
the transition to smaller scale is fraught with
scientific and economic problems. Some stem from
the highly variable nature of contaminant
distributions at hazardous waste sites, and other
problems result from the limitations of available
methods, tools, and theories.
Even though a given parameter, such as hydraulic
conductivity, can be measured correctly and with
great reproducibility, it is difficult to know how
closely an observation actually represents the
vertical and horizontal distribution of conductivities
found at a site. Therefore, when using a conceptual
model to interpret contaminant transport processes,
it is crucial that special attention be given to the
spatial and temporal variations of the collected data.
To circumvent the large numbers of measurements
and samples needed to reduce uncertainties in
dealing with subsurface parameters, more compre-
hensive theories are constantly under development.
The use of these newly developed theories, however,
is often frustrating because many call for data which
are not yet practically obtainable, such as chemical
interaction coefficients or relative permeabilities of
immiscible solvents and water. Therefore, modern
contaminant transport and fate studies necessarily
involve a compromise between sophisticated
theories, current limitations for acquiring data, and
economics.
Remediation
A major issue in cleaning up ground-water
contamination is determining when remediation is
complete. In remedial actions, the level of
contamination measured at monitoring wells may be
dramatically reduced after a moderate period of
time, but low levels of contamination usually persist.
In parallel, the contaminant load removed by
extraction wells, for instance, declines over time and
gradually approaches a residual level in the latter
stages (Figure 1). A decision must be made whether
to continue or end remediation. By continuing
remediation, efforts will be made to clean up small
amounts of residual contamination. However, if
remediation is ended prematurely, an increase in the
level of ground-water contamination may follow
(Figure 2).
There are several contaminant transport processes
that may be responsible for the persistence of
residual contamination and the kind of effect
depicted in Figure 2. In order to generate such an
ON
OFF
MAX
o
I
ui
o
o
RESIDUAL
CONTAMINATION
\
— TIME -^~
Figure 1. Apparent cleanup by pump-and-treat remediation.
OFF
MAX
O
i=
m
O
o
o
— TIME
1 2
Figure 2. Contaminant rebound after remediation
ceases.
effect, releases of contaminant residuals must be
slow relative to water movement through the
subsurface caused by pumping. Transport processes
that generate this kind of behavior include: (1)
diffusion of contaminants within spatially variable
sediments; (2) hydrodynamic isolation; (3) sorption-
desorption; and (4) liquid-liquid partitioning.
Flow through the zones of highest hydraulic
conductivity results in rapid cleansing of these zones
by extraction wellfields, but cleanup of contaminants
in low permeability zones can only occur after the
slow process of diffusion takes place (Figure 3). The
situation is similar, though reversed, for in situ
remediations that require the injection and delivery
of nutrients or reactants to the zone of intended
action. Because the surface area of low-permeability
sediments is greater than that for high-permeability
sediments, greater amounts of contaminants accum-
ulate on them. Hence, the majority of contaminant
reserves may be available only under diffusion-
controlled conditions in many heterogeneous
settings.
-------
LOW PERMEABILITY
STRATA I
ADVECTION
Figure 3. Remediation effectiveness limited by variations
in permeability.
The operation of any remediation wellfield results in
the formation of stagnation zones downgradient of
extraction wells and upgradient of injection wells.
These zones are hydrodynamically isolated so that
mass transport within the isolated water will take
place only by diffusion. Because of this, the portion of
the contaminant plurne lying within a wellfield's
associated stagnation zones will not be effectively
remediated.
For remediation efforts involving compounds that
readily sorb to aquifer materials, the number of pore
volumes to be removed depends not only on the
sorptive tendencies of the contaminant, but also on
whether flow rates during remediation are too rapid
to allow contaminant levels to approach local
equilibrium (Figure 4). If insufficient contact time is
allowed, the affected water is advected away from
sorbed contaminants prior to reaching equilibrium
and is replaced by upgradient fresh water. This
method of removal generates large volumes of mildly
contaminated water where small volumes of highly
contaminated water would otherwise result.
When non-aqueous phase liquid (NAPL) residuals,
such as gasoline, are trapped in pores by surface
tension, diffusive liquid-liquid partitioning controls
dissolution of the toxic compounds within the NAPLs
into the ground water. As with sorbing compounds,
flow rates during remediation may be too rapid to
allow saturation levels of the partitioned contam-
inants to be reached (Figure 5) and large volumes of
mildly contaminated water will be generated.
The practical use of remediation wellfields and other
ground-water cleanup technologies are highly
dependent on site-specific knowledge of the influence
of transport processes on contaminant levels. There
is still much to be learned about highly specific and
cost-effective remediations; however, far more could
be accomplished if the processes that govern the
behavior and treatability of contaminants were
actively investigated at each site. In general, con-
^ORGANIC CARBON OR
MINERAL OXIDE SURFACE
o
UJ
O
o
o
EQUIL, CONG.
SLOW
DESORPTION
INITIAL RAPID
DESORPTION
TIME
Figure 4. Sorption limitations on remediation effectiveness.
ventional field characterization efforts have not led
to satisfactory remediations. Recent transport-
process-oriented approaches of characterization are
resulting in more permanent and cost-effective
remediations.
There are many misconceptions regarding the
processes affecting the transport and fate of
contaminants in the subsurface. Some of these are
relatively easy to address by educational efforts
while others can be addressed only by applied
research. This document will describe some of the
information known about transport and fate of
contaminants in the subsurface. By understanding
and using this information, as well as information
derived from future research, the ground-water
protection objectives required by Congress may be
met.
EPA's Transport and Fate Research
Program
The U.S. EPA Office of Research and Development
(ORD) operates 12 national laboratories, several of
which address various aspects of ground-water
contamination. The Environmental Monitoring
Systems Laboratory in Las Vegas, Nevada develops
-------
ADVECTION
UQUID:LIQUID
PARTITIONING
SOLUBILITY
LIMITED
GROUND-WATER VELOCITY »~
Figure 5. Liquid partitioning limitations on remediation
effectiveness.
monitoring systems, with emphasis on hazardous
waste site characterization, underground storage
tank leak detection, and soil vapor monitoring.
Engineering aspects of the prevention and
remediation of ground-water contamination are
addressed by the Risk Reduction Engineering
Research Laboratory in Cincinnati, Ohio which
develops source control measures (e.g., landfill
covers and liners, and waste caps) and waste disposal
techniques (e.g., solidification, pump-and-treat, and
incineration). Other ORD laboratories also support
research that is germane to ground-water contam-
ination issues (notably, those laboratories in Ada,
Oklahoma; Athens, Georgia; Edison, New Jersey;
and Corvallis, Oregon).
The Robert S. Kerr Environmental Research
Laboratory (RSKERL) in Ada, Oklahoma focuses on
these issues as its primary charge. The Kerr
Laboratory serves as EPA's center for studies of the
transport and fate of contaminants in the subsurface.
The research program includes development of
methodologies for protection and restoration of
ground-water quality, and evaluation of the applica-
• bility and limitations of using natural soil and
subsurface processes for the treatment of hazardous
wastes. RSKERL's efforts in the 1980s have
increasingly focused on improvement of site charac-
terization and remediation methods, with special
emphasis on identification and quantification of the
mechanisms by which natural processes govern the
transport and fate of contaminants in the subsurface.
Transport processes research at RSKERL is divided
into three major areas:
1. Hydrologic processes, which act to influence
the movement of water (the primary vehicle
for subsurface contaminant movement).
2. Abiotic processes, which are the physical and
chemical interactions that cause contami-
nants to move at different rates than those of
the ground water.
3. Biotic processes, which are the microbially
mediated transformations of contaminants
in the subsurface to other compounds.
In the subsurface, however, these processes are
inseparable, and RSKERL's comprehensive research
goal is to ultimately have the knowledge to integrate
the influences of these processes into a singular
understanding of contaminant behavior in the sub-
surface.
-------
CHAPTER 2
PHYSICAL PROCESSES CONTROLLING THE TRANSPORT OF
CONTAMINANTS IN THE AQUEOUS PHASE
Carl D. Palmer and Richard L. Johnson
introduction
Interest in the transport and fate of contaminants in
terrestrial subsurface environments is based on
concern for the protection and remediation of both
ground- and surface-water resources. To achieve this
protection, it is necessary to: (1) predict.the time of
arrival and concentration of contaminants at a
receptor such as a monitoring well, a water-supply
well, or a surface water body; (2) design safe, cost-
effective waste facilities; (3) install effective monitor-
ing systems; and (4) develop efficient and cost-
effective strategies for remediation of contaminated
aquifers. To attain these goals, the processes
involved in the transport and transformation of
contaminants in both porous and fractured media,
and under either saturated or undersaturated
conditions must be understood. This chapter will
discuss some issues associated with the physical
processes controlling contaminant transport in the
subsurface. The advection-dispersion theory in
saturated, porous media will be described as well as
the issues pertaining to contaminant transport in the
vadose zone and in fractured rock. This information
can assist contractors, consulting engineers, and
scientists in designing more efficient and cost-
effective monitoring networks and remediation
strategies, and safer waste facilities.
Contaminant Transport in Saturated,
Porous Media
Advection-Dispersion Theory
If the concentration of a contaminant were measured
in a monitoring well located between a contaminant
source and a receptor such as a water supply well
(Figure 6), a concentration breakthrough curve could
be obtained (Figure 7). In the graph, the
concentrations do not immediately increase in a step
function but, instead, increase more gently in an S-
shaped curve. In a one-dimensional, homogeneous
system, the arrival of the center of mass is the result
of advection while the spread of the breakthrough
curve is the result of dispersion.
Advection is the transport of a non-reactive,
conservative tracer at an average ground-water
velocity. The average linear velocity, v, at which
ground water flows through a porous aquifer is:
v = - (K/6f) (dh/dx)
t>
(1)
where K is the hydraulic conductivity of the
formation in the direction of ground-water flow, 6t is
the porosity of the formation, and (dh/dx) is the
hydraulic gradient in the direction of ground-water
flow (Freeze and Cherry, 1979). The velocity given by
this equation can be substantially different for
solutes that react through precipitation/dissolution,
adsorption, and/or partitioning within the geologic
media (see Chapter 5).
The study of dispersion phenomena is important for
predicting the time when an action limit, a
concentration limit used in regulations such as
drinking water standards, will be reached and for
determining optimal, cost-effective strategies for
aquifer remediation. The classical mathematical
approach used to determine solute transport in
porous media is the advection-dispersion equation.
This equation is written in its one-dimensional form
as:
-------
WASTE
AQUIFER
VADOSE
, ZONE \
MONITORING
WELL
WATER
SUPPLY
WELL
Figure 6. Site containing a monitoring well, contaminant source, and receptor.
BREAKTHROUGH CURVE
1.0
0.5
o
o
0.0
TIME
Figure 7. Concentration of a contaminant in a monitoring well.
-------
3 / acy ac ac
— D — - v — = —-
ax\ ax / .ax at
(2)
where D is the dispersion coefficient, v is the ground-
water velocity, C is the concentration of the solute, x
is the spatial coordinate, and t is time. This equation
is an expression of the mass balance of a contami-
nant within the aquifer as a result of dispersion,
advection, and change in storage. These processes
are represented respectively by the first, second, and
third terms of Equation 2.
Early studies of the dispersion coefficient revealed
that it varies with the ground-water velocity
(Perkins and Johnston, 1963). A plot of the D/D0
versus vd/D0 (Figure 8) where D0 is the molecular
diffusion coefficient of the contaminant, v is the
ground-water velocity, and d is the grain diameter of
the porous medium, shows that the dispersion
coefficient is relatively constant at low velocities, but
increases linearly with velocity as the ground-water
velocities increase. These experimental results led
investigators to propose that the dispersion
coefficient can be written as the sum of two terms: an
effective molecular diffusion coefficient, D^, and
mechanical dispersion coefficient, Dm:
D = D, + D
d m
(3)
The mechanical dispersion coefficient is proportional
to the velocity:
D = av
m
(4)
where a is the constant of proportionality and is
known as the dispersivity parameter. This param-
eter continues to be the focus of a great deal of
research and controversy. At a small scale of meas-
urement, mechanical dispersion is the result of: (1)
velocity variations within a pore; (2) different pore
geometries; and (3) the divergence of flow lines
around the sand grains present in a porous medium
(Gillham and Cherry, 1982).
The effective molecular diffusion coefficient is a
value for diffusion within the porous medium. It can
be obtained from the solution diffusion coefficient,
Do, by:
D, =
d
(5)
where T is the tortuosity of the medium. Tortuosity is
a factor that accounts for the increased distance a
10
10
10
-1
10
I I I I Mill I I I I I Illl I I I 1 I 11II
DL = Longitudinal Dispersion Coefficient
Do = Molecular Diffusion Coefficient
v = Solute Velocity
d = Average Grain Diameter
I I I I I Illl
I I I I/I I
TRANSITION ZONE
DIFFUSION
DOMINATED
L= D0 T
i i i i mi
I I I Illl
i i i 1111 i i i i mi i i i i i i in
ADVECTION
DOMINATED
10
-3
10
-2
10
-1
10
1
10
10
vd/D,
Figure 8.
Dispersion coefficient as a function of ground-water velocity, v, and grain diameter, d (after Perkins and Johnston.1963).
-------
diffusing ion must travel to get around the sand
grains. For granular, porous media, x is typically in
the range of 0.6 to 0.7 (Perkins and Johnston, 1963;
Gillham, et al., 1984). A simple theoretical calcu-
lation indicates that, rather than travelling through
a^grain of diameter d, the solute must travel the
distance around the grain or nd/2. Thus, the
tortuosity should have a value of approximately 0.64.
The advection-dispersion equation (Equation 2) can
be solved with appropriate boundary and initial
conditions to yield the concentration breakthrough
curves or concentration distribution curves. For a
one-dimensional system with a constant concen-
tration of solute at the boundary, the concentration
at a given distance from the source results in an S-
shaped curve (Figure 9). The concentration break-
through curves arrive at points further from the
source at later times and are more spread out. The
concentration distribution curves look like a mirror
image of the breakthrough curves with greater
concentrations toward the source (Figure 10).
The discussion of the advection-dispersion equation
has been limited thus far to the one-dimensional
case. The same general principles can be applied to
two- and three-dimensional problems. Figure 11
shows the transport of a contaminant slug through a
porous aquifer in two dimensions. As the center of
mass of the slug moves further from its initial
location within the aquifer by advection with the
ground-water flow, the slug spreads. The spreading
in the direction of ground-water flow is the longi-
tudinal dispersion, while the spread in the direction
perpendicular to the ground-water flow is known as
the transverse dispersion.
If the slug is viewed in three dimensions, there are
three dispersion coefficients, one longitudinal and'
two transverse. In its general mathematical-,
formulation, the dispersion coefficient is a second
rank tensor. More mathematically detailed descrip-
tions of the advection-dispersion equation can' be ,'•
found in the works of Bear (1979 and 1969). •
Application of Advection-Dispersion
Theory
One-, two-, and three-dimensional advection dis-
persion equations have been used to simulate the
transport of contaminants. However, discrepancies
between theory and laboratory experiments were
observed. Investigators attribute these discrepancies
to a variety of mechanisms including immobile zones
of water within experimental columns, solution-solid
BREAKTHROUGH CURVE
1
o.s
LU
0.4
ceo
O 0.2
O
DISPERSION
TIME
Figure 9. Concentration breakthrough curves derived from the advection-dispersion equation.
-------
CONCENTRATION DISTRIBUTION
LU
LU
DC
DISTANCE
Figure 10. Concentration distribution curves derived from the advection-dispersion equation.
interface processes, anion exclusion, and diffusion in
and out of aggregates. An apparent scale dependency
of the dispersion coefficient also has been observed
(Table 1). Laboratory scale experiments generally
Table 1. Longitudinal Dispersivity Values Obtained from
Different Scale Experiments (After Gillham and
Cherry, 1982)
Type of Test Longitudinal Dispersivity
Laboratory Tests
Natural Gradient Tracer Tests
Single Well Tests
Radial and Two Well Tests
Model Calibration of Contaminant
Plumes '
0.0001 to 0.01
0.01 to 2 m
0.03 to 0.3 m
0.5 to 15 m
3 to 1 00 m
m
yield small values of longitudinal dispersivity
(0.0001 to 0.01 m) while field tracer tests and model
calibration of contaminant plumes yield longitudinal
dispersivity values in the 10 to 100 m range. Even
within a single tracer experiment, longitudinal
dispersivity values are observed to increase with
increasing transport distance (Figure 12). This
apparent increase in the longitudinal dispersivity
parameter with an increasing spatial or time scale
indicates that the assumptions often used when
applying classic advection-dispersion theory in
natural geologic materials are not applicable.
Furthermore, longitudinal dispersivity values gen-
erally were considered to be only 10 to 30 times
larger than transverse dispersivity values. If trans-
verse dispersivity values are large, contaminant
plumes will spread over the entire thickness of an
aquifer (Figure 13A). This is contrary to the long,
thin plumes (Figure 13B) observed in the field where
detailed three-dimensional monitoring was per-
formed (MacFarland, et al., 1983; Kimmel and
Braids, 1980). These observations indicate that
transverse dispersivity must necessarily be very
small and in many situations may even be close to
zero (Sudicky, 1986; Frind and Hokkanen, 1987). If
so, the transverse dispersion coefficient would be
equal to the effective diffusion coefficient within the
medium.
There is a growing consensus among contaminant
hydrologists that the large longitudinal dispersion
-------
ADVECTION AND DISPERSION
OF A CONTAMINANT SLUG
cc
LLJ
o
LL
O
CC
o
Figure 11. Transport pf a contaminant slug through a porous aquifer.
X2
A. HYPOTHETICAL CONTAMINANT PLUME
WITH A LARGE TRANSVERSE DISPERSIVITY
•~
WASTE
to
0.5
"
0.3
0.2
0.1
0
LONGITUDINAL DISPERSIVITY
DATA FROM FHEYBERQ (1986)
B. HYPOTHETICAL CONTAMINANT PLUME
WITH A SMALL TRANSVERSE DISPERSIVITY
WASTE
20 40 60 80
DISTANCE FROM SOURCE (m)
100
Figure 12. Increase in longitudinal dispersivity
with transport distance.
le
figure 13. Hypothetical contaminant plumes for
large (A) and small (B) dispersivities.
10
-------
coefficients observed in field experiments and ob-
tained by model calibration of contaminant plumes
are the result of aquifer heterogeneity. The effects of
heterogeneity on the spread of contaminants is easily
illustrated in an ideally stratified aquifer with layers
of sediment of different hydraulic conductivities
(Mercado, 1967). Contaminants move rapidly along
the layers with higher permeability and more slowly
along the lower permeability layers (Figure 14). If
water is sampled from monitoring wells that are
screened through the various layers, the sample is an
integration of the concentrations in each layer.
Plotting concentration versus distance from the
source reveals a curve with a large amount of spread
even though only advection is considered. This
macroscopic dispersion is the result of aquifer hetero-
geneity and not the pore-scale processes described
above.
C0 ..*.
t
'
• f
\
~t
• 1
• •
ft
Ki
Ka
Ks
"Ki|
Ks
—
—
o
o
DISTANCE
Figure 14.
Contaminant distributions and concentrations in an
ideally stratified aquifer (after Gillham and Cherry,1982).
Investigators can calculate contaminant transport
rates through stratified aquifers by assuming that
hydraulic conductivities follow a Gaussian distribu-
tion. When completed, these analyses indicate that
the spread in the contaminant distribution is
proportional to the mean distance travelled, L,
rather than L1/2 which is obtained for the classic
advection-dispersion theory (Mercado, 1967). Thus,
the spatial spread of the contaminant concentrations
is greater than that predicted by the advection-
dispersion theory. If the stratified medium is
assumed to be homogeneous and if the advection-
dispersion theory is applied, the apparent disper-
sivity parameters increase with distance from the
source.
All geologic formations are heterogeneous and none
are perfectly stratified. Bedforms are often lenticular
and may contain cross-stratification or graded bed-
ding. Detailing the distribution of hydraulic conduc-
tivity in the subsurface is a formidable task that
cannot be done at waste sites on a routine basis.
Heterogeneity can, nonetheless, be quantified by
considering hydraulic conductivity to be a random
process, and determining its important statistics
such as the mean, variance, and autocorrelation
function.
Such statistical analyses were performed on the
Borden aquifer (Ontario, Canada) where perme-
ameter measurements on more than 3,000 samples
were analyzed (Sudicky, 1986). In an aquifer that is
ostensibly described as homogeneous, hydraulic
conductivity was found to vary by more than two
orders of magnitude. Statistical analysis of this data
further revealed that the 'hydraulic conductivity is
log-normally distributed. Autocorrelation length-
scales are 1.6 m in the horizontal direction and 0,10
m in the vertical direction. Using this information
and applying the stochastic theory of Gelhar and
Axness (1983), Sudicky (1986) calculated the
asymptotic longitudinal dispersivity for the Borden
aquifer to be 0.61 m, a value that is close to the 0.43
m reported by Freyberg (1986) for the large-scale
tracer test conducted at the site The asymptotic
transverse dispersivity values calculated by Sudicky
(1986) are very close to zero, indicating that the
transverse dispersion coefficient is on the order of the
effective molecular diffusion coefficient.
Gillham, et al. (1984) proposed an advection-
diffusion model for the transport of solutes in
heterogeneous, porous media. This model recognizes
that aquifers are heterogeneous and that advection
is the key process controlling the rate of movement of
solutes through the layers of higher permeability. As
the front moves through those layers, some of the
solute is lost to the lower permeability layers via
molecular diffusion. This advection-diffusion model
was tested in the laboratory (Sudicky, et al., 1985)
using a sandbox model composed of a layer of sand
sandwiched between layers of silt. If molecular
diffusion in the transverse direction is ignored, the
one-dimensional advection-dispersion equation can
be applied (the dashed line in Figure 15). The
experimental data, however, revealed retardation in
the solute front and a much greater spread in the
breakthrough curve. A mathematical solution that
includes diffusion into the lower-permeability layers
was able to simulate the experimental data. If the
one-dimensional equation is used to analyze the
experimental curves, a large apparent dispersivity
would be obtained for the higher permeability layer.
While aquifer heterogeneity is a major factor
contributing to the spread of contaminants, other
processes contribute to contaminant transport
characteristics as well. Diverging flow lines spread
contaminants by advection over a larger cross-
section of the aquifer. Temporal variations in the
water table can change the direction of ground-water
11
-------
o
<
DC
ill
O
O
o
BREAKTHROUGH CURVES
SHOWING EFFECT OF TRANSVERSE DIFFUSION
0.5
i
i
i
i
i
i
i
i
i
10
i 1 , ,
ADVECTION-DISPERS10N
ADVECTION-DIFFUSION
TRAVEL DISTANCE * 1.0 m
SAND THICKNESS =» 0.03 m
GROUNDWATER VELOCITY - 0.10 m/day
DISPERSIVITY - 0.001 m
DIFFUSION COEFFICIENT « 1.2 X 10'*ma/»
20 30
TIME (DAYS)
40
50
Figure 15. Breakthrough curves showing the effect of transverse diffusion.
flow and contribute to the lateral spread of contam-
inants. Also, variations in the contaminant concen-
trations at the source can cause apparent dispersion
in the longitudinal direction (Frind and Hokkanen,
1987).
Apparent spreading in contaminant plumes also may
be the result of ground-water sampling methods.
Insufficient well purging may result in under-
estimation of contaminant concentrations at select
locations within the aquifer. Monitoring wells with
different screen lengths integrate ground water from
disparate portions of the aquifer and may yield
dissimilar contaminant concentrations.
Dispersion is a phenomenon that often is used as a
mathematical convenience to correct for ignorance
about the aquifer's heterogenous nature and a poor
understanding of the processes occurring within the
aquifer. As the fundamental processes controlling
the distribution of contaminants in the subsurface
are understood, better and more cost-effective tech-
niques for aquifer remediation will be designed.
Diffusive Transport Through Low
Permeability Materials
Unfractured clays and rocks often have hydraulic
conductivities less than 1O9 m/s. A review of ground-
water flow through such low-permeability forma-
tions is provided by Neuzil (1986). However, for such
materials, the diffusive transport of contaminants is
large compared to advective transport.
The diffusion of solutes through porous media has,
been studied in the laboratory (Gillham, et al., 1984;
Robin, et al., 1987). Experiments confirm the appli-
cability of Fickian diffusion models and Equation 5
for calculating the effective diffusion coefficient.
Tortuosity factors in sand-clay mixtures were in the
range of 0.59 to 0.84 for 36C1 and 0.33 to 0.70 for
tritium (Gillham, 1984).
A study of fine-grained Quaternary deposits in
southern Ontario by Desaulniers, et al. (1981, 1986)
established that the movement of Cl" and !8O
through the deposits occurred through molecular
diffusion. The transport of inorganic ions through a
clay beneath a municipal landfill also was shown to
be predominantly by molecular diffusion (Goodall
and Quigley, 1977; Crooks and Quigley, 1984).
Similarly, a study of the transport of organic contam-
inants through a saturated clay beneath a hazardous
waste site shows that the distribution of contam-
inants is controlled by Fickian diffusion (Johnson, et
al., 1989).
Contaminants can contaminate aquifers by diffusing
across natural aquitards or clay liners. The extent of
ground-water contamination will depend on the
diffusive flux, rate of ground-water flow in the
12
-------
aquifer, and length of the source area in the direction
of ground-water flow. Diffusion of organic contam-
inants is discussed in more detail in Chapter 6. .
Effects of Density
The discussion so far has been limited to cases where
the contaminant plume does not have a density
significantly different from the native ground water.
In some cases, dissolved concentrations are large
enoiigh -that the density of the contaminant plume
may contribute to the direction of solute transport.
The contribution of density to the vertical component
of ground-water velocity, Vg, can be calculated using
the concept of equivalent freshrwater head (Frind,
1982) by:
V =
•V
(6)
where KZ2 is the hydraulic conductivity in the
vertical direction, 0t is porosity, p is the density of
the contaminated water, and p0 is the density of the
native ground water.
As an example, assume that the density of ground
water within an .aquifer is l.OQ, the natural
horizontal gradient is 0.005, and the natural vertical
gradient is 0.000. If the density of the contaminated
water is equal to the density of the native ground
water, the contaminant plume moves horizontally
under-the naturally existing hydraulic gradient. If
the density of the contaminated water is 1.005, then
the driving force in the vertical direction is the same
as the driving force in the horizontal direction. If the
aquifer is isotropic, then the resultant vector of these
twd forces plunges at 45° into the aquifer. Under
these conditions, a contaminant plume moves deeply
into the aquifer and may not be detected with shal-
low monitoring systems installed under the assump-
tion of horizontal flow. The density of seawater,
which contains. 36,000 mg/L total dissolved solids, is
1.025 and the density of pure water is close to 1.000.
Therefore, the density of 1.005 corresponds to
approximately 7,000 mg/L.
Retardation of Contaminants
Not-all:, solutes are transported through- geologic
material at :the same rate. If solutes undergo chem-
ical reactions while being transported, their rate of
movement may be substantially less than the aver-
age' 'rate.of ground-water flow. Such chemical
reactions include precipitation, adsorption, ion ex-
change,, and partitioning into soil organic matter or
organic solvents. While these topics are discussed in
more detail in Chapter 5, the effect of this
retardation on the breakthrough curves is intro-
duced here.
One simple form of the differential equation for
contaminant transport with retardation is:
a / D ac \ v ac ac
_
dx \ R ax / R
(7)
where R is a constant known as the retardation
factor and the other parameters are as defined above.
If the retardation factor is equal to 1.0, the solute is
nonreactive and Equation 2 is obtained. If R is
greater than 1.0, the average velocity of the solute,
v/R, is less than the velocity of the ground water and
the dispersion of the solute, D/R, is likewise reduced.
If a monitoring well is located at such a distance from
a contaminant source that it takes time ti for a
nonreactive solute to travel from the source to the
well, it will take 2ti for a contaminant with a
retardation factor of 2 to reach that same well and
4ti for a contaminant with a retardation factor of 4
(Figure 16).
<
CC
o
u
1.0
0.5
0.0
R= 1
TIME
Figure 16. Time required for movement of contaminants
at different retardation factors.
Contaminants with lower retardation factors are
transported greater distances over a given time
period than contaminants with larger retardation
factors (Figure 17). A monitoring well network has a
greater chance of encountering contaminants with
low retardation factors simply because they occupy a
greater volume of the aquifer. Thus, estimates of the
total mass of a contaminant with a retardation factor
of 1.0 may be more accurate than those for contam-
inants with greater amounts of retardation. Also,
estimates of the time to remove nonreactive contam-
inants may, therefore, be more accurate than those
time estimates for retarded contaminants. This is
particularly important because the slow movement
of retarded contaminants may control .the time and
cost necessary to completely remediate a site.
13
-------
RETARDATION AND MONITORING
1,2,3 1 & 2
WASTE DETECTED DETECTED
1 ONLY
DETECTED
Figure 17. Transport of contaminants with lower retardation factors at a waste site.
Flow and Transport in the Vadose Zone
The vadose zone is that portion of the soil between
the ground surface and the water table and includes
the capillary fringe. More generally, the zone is
defined as that region in which the pressure head is
less than zero. Because it is part of the overall flow
path, this zone can be very important to the trans-
port of contaminants. Its length and the velocity of
the contaminants passing through it, therefore,
should be included in estimates of transport times.
The vadose zone often contains greater amounts of
organic matter and metal oxides than the saturated
zone. Contaminants can adsorb onto these materials,
making their rate of movement substantially less
than in the saturated zone. Further, materials
adhering to these adsorbents can act as a source of
contaminants to the saturated zone even after reme-
diation. In addition, the activity of microorganisms
in the vadose zone generally is considered to be much
greater than below the water table. Finally, the
unsaturated portion of the vadose zone can be a
pathway for the transport of gases and volatile
organics. These characteristics of the vadose zone
can be important when predicting the transport of
contaminants from a waste site and designing
systems for remediation.
Transport of Water and Solutes
The flow of water through the vadose zone can be
described by a differential equation that is analogous
to the ground-water flow equation. The equation's
one-dimensional form is:
3 / dip \ d dip
- K(ip) — + - [K(ip)] = T (ip) — (8)
dz \ dz / dz dt
where K(ip) is the hydraulic conductivity, F(ip) is the
specific water capacity (d9w/dip), 9W is the volumetric
water content, and ip is the soil water pressure head.
The key difference between this equation and the
equation for one-dimensional ground-water flow is
the dependence of the hydraulic conductivity on ip
and hence on 9W. This makes Equation 8 nonlinear
and more difficult to solve than the ground-water
flow equation. If the hydraulic conductivity is
constant and if the total head (h) is given as the sum
of the pressure head and the elevation head (z), then
Equation 8 simplifies to the one-dimensional ground-
water flow equation.
The pressure head and the volumetric water content
are related and often are plotted as a characteristic
14
-------
curve (Figure 18). If a saturated soil with zero
pressure head is drained, ip decreases while the
volumetric water content remains constant and
equal to the porosity of the soil. The volumetric
water content is maintained at this level until
sufficient negative pressure (the air entry value) is
achieved to allow air to begin to enter the soil. At
this point, the volumetric water content decreases in
response to the decreasing pressure head. At low
pressure head, the curve begins to level off and
asymptotically approaches the residual water
content of the soil.
CHARACTERISTIC CURVE
Air Entry Valua
,,0.5
MAIN
DRAINAGE
CURVE
-300
-200 -100
PRESSURE HEAD (CM OF WATER)
Figure 18 Volumetric water content versus pressure head.
If water is added to the drained soil, the plot of
volumetric water content versus pressure head does
not follow the main drainage curve described above
but instead follows another path known as the main
wetting curve. Thus, the relationship7between
pressure head and volumetric water content is not
unique. If a soil is completely drained and then
saturated again, the volumetric water content would
initially plot along the main drainage curve and then
along the main wetting curve. If the draining or
wetting process is interrupted before the cycle is
complete, the data plots between the two main
curves along the primary scanning lines (Figure 19).
If the wetting or drainage of the soil is reversed while
the data is on the primary scanning lines, the data
plots along yet a different path (secondary scanning
lines). The dependence of volumetric water content
(or any other soil property) at a given pressure head
on the wetting and drying history of the soil is known
as hysteresis.
If the hydraulic conductivity is a function of the
pressure head, it also must be a function of the volu-
metric water content. For instance, the hydraulic
conductivity of a soil may decrease by more than two
orders of magnitude as the volumetric water content
decreases from saturation to residual water content.
CHARACTERISTIC CURVE
SCAN LINES
-300
-200
PRESSURE HEAD (CM OF WATER)
Figure 19. Volumetric water content versus pressure head
showing primary scanning lines.
Solute transport in the vadose zone can be described
by an advection-dispersion equation with an one-
dimensional form of:
-60 —
w 3z
a(gO
3z
at
(9)
where C is the solute concentration, D is the vadose
zone dispersion coefficient, Ow is the volumetric
water content, and q is the volumetric water flux.
The dispersion coefficient has been assumed to be
analogous to the dispersion term in the saturated
zone:
D = D t + av (0 )
o w
(10)
where v(6w) is the solute velocity and is equal to
q/9w. Recent experiments by Bond (1986) demon-
strate that Equation 10 is the correct form of the
dispersion coefficient for transport in the vadose
zone.
The application of Equations 9 and 10 to field
situations is plagued by at least as many problems as
those discussed earlier for Equation 2. Heterogeneity
in the vadose zone may be the result of soil structure
(aggregates) or macroscopic pores such as earthworm
holes, decayed root channels, animal burrows, and
fractures, all of which can substantially alter the
flow of water and the transport of solutes through the
vadose zone (White, 1985). Van Genuchten and Jury
(1987) reviewed several modeling approaches being
developed to investigate these situations.
Understanding the processes that control the
movement of water and solutes in the vadose zone
can provide insight into field observations of
contaminant distributions and can be used to design
15
-------
storage facilities for wastes. Gillham (1984) used the
general concepts of flow in the vadose zone to explain
the role of the capillary fringe in the apparently
disproportionate rise in shallow water tables with
small amounts of recharge. Such water table changes
help to explain stream-flow generation (Abdul and
Gillham, 1984). Further, such a theory suggests that
the characteristics of the capillary fringe and the
vadose zone may contribute to the spreading of
contaminants below the water table. Finally, Frind,
et al. (1976) suggest that the differences in the
unsaturated hydraulic conductivity between coarse-
and fine-grained materials can be utilized in waste
storage facilities.
Transport in the Gas Phase
Transport of gases and contaminants through the
unsaturated zone can be an important consideration
in certain field situations. Some organic contam-
inants are volatile and can partition from the liquid
phase into the vapor phase. These vapors are
transported through the unsaturated zone and
eventually may diffuse into the atmosphere. The key
physical processes that affect the transport of gases
in the vadose zone are diffusion and advection, with
diffusion playing the largest role. This is the result of
the large diffusion coefficient for gases (10-5 m2/s)
compared to solutes (10-9 m2/s). Many volatile
organic chemicals have equilibrium concentrations
that are high enough to increase the density of the
vapor phase to 1.5 g/cm3. This high density, in
principal, should cause these vapors to sink to the
capillary fringe. Cultural features, such as parking
lots, streets, and foundations, can limit the exchange
of gases with the atmosphere. The transport of
volatile organics through the soil-gas phase also will
be affected by the partitioning of the gas phase into
the soil water, adsorption, and biodegradation
(Johnson and Pankow, 1987).
There is an interest in using vapor monitoring wells
to locate contaminant plumes in the saturated zone
because these wells are much less expensive to
install than standard ground-water monitoring
wells. This concept is based on the premise that
volatile organics within a contaminant plume
located just below the water table will partition into
the vapor phase in the overlying unsaturated zone
where they then can be detected. In reality, the
transport of volatile organics through the un-
saturated zone is complicated by the chemical and
physical processes discussed above and the general
heterogeneous nature of soils. Vapors may be
transported along high permeability layers to points
distant from the ground-water source. Partitioning,
adsorption, and biodegradation may reduce concen-
trations of contaminants in the vapor phase to levels
that cannot be detected. Therefore, the use of vapor
sampling to establish the existence of contaminant
plumes in the subsurface is not always a reliable
technology.
The volatility of certain organic contaminants can be
exploited for the purpose of remediation. If there are
residual solvents or petroleum products in the un-.
saturated zone, it is possible to remove the volatile
fraction through vapor pumping. Wells are installed
in the unsaturated zone and soil air extracted
through a vacuum system. As the pressure drops and
clean air passes through the soil, the organic contam-
inants partition into the vapor phase where they are
then extracted by the vapor pumping well. The use of
such systems is becoming more commonplace, and
more efficient methods for their application are
being developed.
In addition to volatile organic contaminants, the
transport of the permanent gases, such as carbon
dioxide and oxygen, also are of great interest. Oxy-
gen can oxidize sulfide minerals (such as pyrite) to
sulfuric acid, which can result in the degradation of
ground-water quality. These problems can be partic-
ularly acute in disturbed lands where mining or
construction has occurred. The presence of carbon
dioxide also can alter ground-water quality by affect-
ing mineral dissolution and the adsorption of metal
ions. Both oxygen and carbon dioxide affect micro-
bial activity and the rate of biodegradation in the
vadose zone.
Contaminant Transport Through
Fractured Media
The models for solute transport discussed up to this
point only address porous media. While such models
are applicable at sites located on recent alluvial
deposits and glacial sediments, they are not neces-
sarily appropriate when designing monitoring sys-
tems or planning remedial activities at waste sites
on fractured rock.
Fractured rock has both primary and secondary
porosity. Primary porosity is the pore space formed
at the time of deposition and diagenesis of the rock
mass. Secondary porosity is the pore space formed as
a result of fracture of the rock.
As in porous media, the transport mechanisms in
fractured media are advection and dispersion. In
fractured rock, however, contaminants are advected
only along the fractures. Dispersion phenomena
within fractured rock is the result of: (1) mixing at
fracture intersections; (2) variations in aperture
across the width of the fracture; (3) variations in
aperture width along stream lines; (4) molecular dif-
fusion into microfractures penetrating the inter-
16
-------
fracture blocks; and (5) molecular diffusion into
interfracture porous matrix blocks. Transport
through fractured media can be described by one of
four general types of models: continuum, discrete
fracture, hybrid, and channel models.
In continuum models, the individual fractures are
ignored and the entire medium is considered to
behave as an equivalent porous medium. These
models may be either single porosity or double
porosity models. Single porosity models are
applicable to fractured crystalline rocks such as
granite and basalt where the only porosity of the
rock mass is the fracture porosity. Double porosity
models assume there is both primary porosity and
secondary porosity. These models are applicable to
media such as sandstones and shales.
Discrete fracture models attempt to describe flow
and transport among individual fractures. These
models require information about each fracture
within the rock mass. The great difficulty of
obtaining this information led to the development of
stochastic models, which utilize information about
the statistical distribution of fracture properties such
as orientation and aperture widths.
Hybrid models are combinations of discrete fracture
and continuum models. An example of this model is
the "Multiple Interacting Continua" (MING) model
(Pruess and Narasimhan, 1985) in which transport
takes place through a three-dimensional fracture
network (the discrete portion) while diffusion into
the interfracture rock matrix (the continuum)
occurs.
Channel models (Tsang and Tsang, 1987) were
developed from both laboratory (Witherspoon, et al.,
1980) and field (Neretnieks, 1985) observations that
the transport of solutes along fractures does not
occur as a uniform front along the width of a
fracture, but in many small fingers or channels. Such
models are only now being investigated.
Studies of fractured rock often make use of the cubic
law for the description of fluid flow and hydraulic
conductivity. The hydraulic conductivity of a
fracture, Kf, is:
K_ = (2brPg/(12u)
(ID
where 2b is the fracture aperture, p is the density of
the fluid, g is the acceleration of gravity, and u is the
dynamic viscosity of the fluid. The above equation
can be modified for application in a continuum
model:
= (2brpgN/(12p.B)
(12)
where B is the thickness of the medium and N is the
number of fractures through that thickness. An
important consideration in the study of fractured
rock is knowing when a fracture network can be
considered to behave like an equivalent porous
media. Numerical studies of single-porosity fracture
networks suggest that such networks behave more
like continua when: (1) fracture density is increased;
(2) apertures are constant rather than statistically
distributed; (3) the orientations of the fractures are
statistically distributed rather than constant; and (4)
larger sample sizes are tested (Long, et al., 1982).
Novakowski, et al. (1985) conducted a tracer test in a
single fracture at about 100 m depth in gneiss at the
Chalk River Nuclear Laboratories in Ontario,
Canada. The fracture aperture of 510 um obtained
from their two-well tracer test does not agree well
with the 60 um aperture obtained from interference
pumping tests. A survey of dispersivities in fissured
rock obtained at various sites in Europe and in the
United States reveals a range of four orders of
magnitude (Neretnieks, 1985). Novakowski, et al.
(1985) obtained a dispersivity of 1.4 m. Tracer tests
at the Oracle, Arizona site are within the range from
several tenths of a meter to a few meters (Cullen, et
al., 1985). As in porous media, the dispersivity
values obtained for fissured media become larger
with increasing length of the flow path (Neretnieks,
1985).
Research into transport phenomena in fractured rock
has been supported primarily by agencies interested
in the potential for transport of radionuclides from
high-level radioactive waste repositories. As a
consequence, the bulk of the research performed has
been of fractured crystalline rock such as granite.
However, many hazardous waste sites are located on
fractured sedimentary rocks such as sandstone and
shale which have primary porosities of 5 to 25
percent in the rock matrix. Fractures increase the
total porosity of these rocks and substantially
increase their hydraulic conductivity, making them
attractive for water supply aquifers. While con-
taminants are advected through these fracture
systems, transport into and out of the porous matrix
is primarily by molecular diffusion (Figure 20). This
latter phenomena is much more important in
fractured, porous rock than in fractured, crystalline
rock.
Fractured, porous aquifers can be found throughout
the conterminous United States. The 20 states with
the greatest potential for ground-water contamina-
tion in fractured, porous media (FPM) are found
primarily in the northeastern United States and
around the Great Lakes (Figure 21). More than 1,500
Comprehensive Environmental Response and Lia-
bility Act (CERCLA) or Superfund sites located on
17
-------
FRACTURED POROUS ROCK
Diffusion
intp Rock
Matrix
Diffusion
Into Rock
Matrix
the rate of movement of the water through the
fractures (vf) by:
= vf/Rf
(13)
where Rf is a retardation factor that accounts for the
loss of contaminant mass from the fracture to the
matrix:
Rf = (b + nB'Vb
(14)
Figure 20. Transport in fractured porous rock.
FPM exist within these 20 states (Johnson and
Pankow, 1987).
The use of porous media models can be appropriate in
evaluating fractured, porous formations if the
concentration of the contaminant can quickly reach
equilibrium with the concentration found in the
fracture (Pankow, et al., 1986). The rate of
movement of a conservative tracer through such a
fractured, porous aquifer (vs) can be calculated from
where b is the half-width of the fracture and B' is the
half-width of the matrix block.
These concepts have been applied to a site at Alkali
Lake, Oregon, and at Bayview Park, Ontario
(Pankow, et al., 1986). At Alkali Lake, the sediments
are highly fractured, the matrix blocks are small (0.3
cm), and the matrix diffusion coefficient is 0.1
cm2/day. In contrast, at Bayview Park, the matrix
blocks are 5 to 35 cm and the matrix diffusion
coefficients are only 0.0032 cm2/day. At Alkali Lake,
the concentrations in the matrix approach equi-
librium with the concentrations in the fractures in
o
Figure 21. States with a high potential for ground-water contamination in fractured, porous rock (bold line outlines EPA Re-
gions).
18
-------
about 5 days, while more than 6,000 days are
required at Bayview Park. As a result, the equiva-
lent porous media model works well at Alkali Lake
but not at Bayview Park. Thus, simple calculations
of the time needed for the matrix block to reach
equilibrium within the fracture can be used as a
guide for the applicability of the equivalent porous
media model for fractured, porous media.
Particle Transport Through Porous
Media
So far, this chapter has only considered the transport
of solutes through porous and fractured media.
However, particles also may be of interest ,to
contaminant hydrologists. The term "particle" can be
used very broadly to include bacteria, viruses,
inorganic precipitates, natural organic matter,
asbestos fibers, or clay.
Particles can be removed from solution by three
major processes: (l) surface filtration; (2) straining;
and (3) physical-chemical processes (McDowell-
Boyer, et al., 1986). Determining which of these
processes is the most effective depends on the size of
the particles (Figure 22). If the particles are larger
than the largest pore diameters, they cannot
penetrate into the porous medium and will be
filtered at the surface of the medium. If the particles
are smaller than the largest pores but larger than
the smallest, the particles are transported into the
porous medium along the larger pore channels.
Eventually, the particles encounter a pore channel
with a smaller diameter and are removed by
straining. If the particles are smaller than the
smallest pore openings, the particles can be
transported great distances through the porous
medium.
The rate at which the particles move through the
porous medium depends on a variety of physico-
chemical processes. For example, the particles may
undergo random collisions with sand grains. A
certain percentage of such collisions result in
particles adhering to the solid matrix (interception).
Chemical conditions also may affect particle trans-
port. For example, if the pH changes, aggregates
may result due to changes in the particles' surface
properties. These larger aggregates then can be
strained or filtered from the water.
Microorganisms are particles that can be transported
through geologic media. The movement of bacteria
and viruses in the subsurface is a significant
problem. More than 50,000 individuals in the United
States suffered from waterborne disease between
1971 and 1979 (Craun, 1984) and about 45 percent of
all reported cases involved ground-water sources. In
addition, increasing interest in the use of
SURFACE
FILTRATION
0O°O0O°
o^o^o^o
0O°O°O0
o^o^o^o
0°0°
STRAINING
cgogogog
O^~\;- ^^O°^~^OQ^~\
oP
o
o
PHYSICAL-
CHEMICAL
O
Figure 22. Filtration mechanisms.
microorganisms for in situ remediation of aquifers
contaminated with organic chemicals necessitates a
greater understanding of the transport and fate of
microorganisms within the subsurface.
There are many processes that limit the movement of
microorganisms through geologic materials. Some
bacteria are large enough to be strained from the
water. In comparison, the smaller viruses can pass
through the pores, but their surfaces are charged
and, like charged ions, may undergo adsorption
under the proper chemical conditions. Like mole-
cules, microorganisms are transported by diffusion.
19
-------
Some microorganisms are motile and move in
response to changes in chemical concentrations. Like
other living organisms, microbes grow and die, and
the rates of these processes must be included in the
description of the transport of microbes in the
subsurface. All of these processes are reviewed in
detail by Yates, et al. (1987) and Matthess and
Pekdeger (1981).
Field tracer tests using baker's yeast (Saccharomyces
cerevisiae) were conducted at a field site in Stanton,
Texas (Wood and Ehrlich, 1978). The baker's yeast
was injected into a sand and gravel aquifer
containing clay lenses. In two separate tests where
bromine and iodine (Br" and D ions were used as
chemical tracers, the baker's yeast arrived at the
monitoring well before the chemical tracers. Wood
and Ehrlich (1978) explain these observations by
suggesting that: (1) the chemical tracers were
retarded because of their adsorption onto the aquifer
material; and (2) the yeast traveled only through
solution channels in caliche deposits within the
aquifer while the chemical tracers flowed through
both the solution channels and the intergranular
pore structure. As the chemical tracer moves
through the larger solution pores, it loses mass to the
smaller pores by molecular diffusion, thereby
retarding the rate of movement of the chemical front.
The larger yeast particles have smaller diffusion
coefficients and are excluded from the smaller pores.
Without the mass loss to the adjacent pores, the
yeast arrives more quickly at the observation well.
Champ and Schroeter (1988) used field tracer tests to
show that non-reactive particle tracers and bacteria
(Escherichia coli) can be rapidly transported through
fractured, crystalline rock. Similar to the results of
Wood and Ehrlich (1978), Champ and Schroeter
(1988) observed that the E. coli arrived before the
bromine tracer. Tracer experiments using native
bacteria and different sizes and types of micro-
spheres were conducted at a Cape Cod site in
Massachusetts (Harvey, et al., 1987). The results
indicate that both size and surface characteristics of
particles affect their movement through the aquifer.
Field evidence of the transport of inorganic particles
in the subsurface was obtained at the Otis Air Force
Base site on Cape Cod in Massachusetts (Gschwend
and Reynolds, 1987). Secondarily treated sewage
containing phosphates was recharged to a sand and
gravel aquifer through rapid infiltration beds over a
30-year period. Downgradient from the rapid
infiltration beds, ground-water samples are found to
contain 100 nm-diameter particles. These particles,
composed of phosphate and iron, may be the mineral
vivianite. The phosphate comes from the recharged
water while the iron is derived from the dissolution
of naturally existing iron within the aquifer.
Another example of particle transport exists at the
Nevada Test Site where particle transport was
identified as a mechanism for the movement of
lanthanide and transition element isotopes from
subsurface nuclear explosion cavities (Buddemeier,
1986).
References
Abdul and Gillham, 1984. "Laboratory Studies of the
Effects of the Capillary Fringe on Streamflow
Generation." Water Resources Research, Vol. 20,
No. 6, pp. 691-698.
Bear, J., 1969. "Hydrodynamic Dispersion." In: Flow
Through Porous Media, R.J.M. De Wiest, Editor.
Academic Press, New York, NY, pp. 109-199.
Bear, J., 1979. Hydraulics ofGroundwater. McGraw-
Hill, New York, NY.
Bond, W. J., 1986. "Velocity-dependent Hydrody-
namic Dispersion During Unsteady, Unsaturated
Soil Water Flow: Experiments." Water Resources
Research, Vol. 22, No. 13, pp. 1881-1889.
Buddemeier, R. W., 1986. "Ground-water Transport
of Radionuclides in Colloidal Form." EOS, Trans-
actions of the American Geophysical Union, Vol. 67,
No. 44, p. 955.
Champ, D. R. and J. Schroeter, 1988. "Bacterial
Transport in Fractured Rock: A Field Scale Tracer
Test at the Chalk River Nuclear Laboratories."
Proceedings of the International Conference on
Waters and Waste Water Microbiology, Inter-
national Association of Water Pollution Research
and Control, Irvine, CA, Feb. 8-11, 1988, pp. 14-1 -
14-7.
Craun, G. F., 1984. "Health Aspects of Ground-water
Pollution." In: Ground-water Pollution Micro-
biology, G. Bitton and C.P. Gerba, Editors. John
Wiley and Sons, New York, NY, pp. 135-179.
Crooks, V. E. and R. M. Quigley, 1984. "Saline
Leachate Migration Through Clay: A Comparative
Laboratory and Field Investigation." Canadian
GeotechnicalJournal, Vol. 21, pp. 349-362.
Cullen, J. J., K. J. Stetzenbach, and E. S. Simpson,
1985. "Field Studies of Solute Transport in Frac-
tured Crystalline Rocks Near Oracle, Arizona." In:
Proceedings, Memoirs of the 17th International
Congress of the International Association of
Hydrologists, Vol. 17, International Association of
Hydrologists, Tucson, AZ, pp. 332-344.
Desaulniers, D. E., J. A. Cherry, and P. Fritz, 1981.
"Origin, Age, and Movement of Pore Water in
Argillaceous Quaternary Deposits at Four Sites in
Southwestern Ontario." Journal of Hydrology, Vol.
50, pp. 231-257.
20
-------
Desaulniers, D. E., etal., 1986. "37C1-35C1 Variations
in a Diffusion Controlled Ground Water System."
Geochimica et Cosmochimica Acta, Vol. 50, pp.
1757-1764.
Freeze, R. A. and J. A. Cherry, 1979. Ground-water.
Prentice-Hall, Inc., Englewood Cliffs, NJ.
Freyberg, D. L., 1986. "A Natural Gradient Experi-
ment on Solute Transport in a Sand Aquifer. 2.
, Spatial Moments and the Advection and Dis-
persion of Nonreactive Tracers." Water Resources
Research, Vol. 22, No. 13, pp. 2031- 2046.
Frind, E. O., R. W. Gillham, and J. F. Pickens, 1976.
"Application of Unsaturated Flow Properties in the
Design of Geologic Environments for Radioactive
Waste Storage Facilities." In: Proceedings, First
International Conference on Finite Elements in
Water Resources, Princeton University, July 1976.
Frind, E. O., 1982. "Simulation of Long-term Tran-
sient Density-dependent Transport in Ground-
water." Advances in Water Resources, Vol. 5, pp.
73-88.
Frind, E. O. and G. E. Hokkanen, 1987. "Simulation
of the Borden Plume Using the Alternating
Direction Galerkin Technique." Water Resources
Research, Vol. 23, No. 5, pp. 918-930.
Gelhar, L. W. and C. L. Axness, 1983. "Three-
Dimensional Stochastic Analysis of Macrodis-
persion in a Stratified Aquifer." Water Resources
Research, Vol. 15, pp. 1387-1397.
Gillham, R. W. and J. A. Cherry, 1982. "Contam-
inant Migration in Saturated Unconsolidated
Geologic Deposits." In: Recent Trends in
Hydrogeology, T. N. Narasimhan, Editor. Geolog-
ical Society of America Special Paper 189, pp. 31-
62.
Gillham, R. A., 1984. "The Capillary Fringe and Its
Effect on Water-table Response." Journal of
Hydrology, Vol. 67, pp. 307-324.
Gillham, R. W., et al., 1984. "An Advection-diffusion
Concept for Solute Transport in Heterogeneous
Unconsolidated Geologic Deposits." Water
Resources Research, Vol. 20 (3), pp. 369-378.
Gillham, R. W., M. J. L. Robin, and D. J.
Dytynyshyn, 1984. "Diffusion of Nonreactive and
Reactive Solutes Through Fine-grained Barrier
Materials." Canadian Geotechnical Journal, Vol.
21, pp. 541-550.
Goodall, D. C. and R. M. Quigley, 1977. "Pollutant
Migration for Two Sanitary Landfill Sites Near
Sarnia, Ontario." Canadian Geotechnical Journal,
Vol. 14, pp. 223-236.
Gschwend, P. M. and M. D. Reynolds, 1987. "Mono-
disperse Ferrous Phosphate Colloids in an Anoxic
Groundwater Plume." Journal of Contaminant
Hydrology, Vol. 1, No. 3, pp. 309-327.
Harvey, R. W., et al., 1987. "Transport of Bacteria
Through A Contaminated Freshwater Aquifer." In:
U.S. Geological Survey Program on Toxic Waste-
Ground-Water Contamination: Proceedings of the
Third Technical Meeting, Pensacola, FL, March
23-27, 1987. U.S. Geological Survey Open-File
Report 87-109, pp. B29-B31.
Johnson, R. L., et al., 1987. "Mass Transfer of
Organics Between Soil, Water, and Vapor Phases:
Implications for Monitoring, Biodegradation, and
Remediation." Proceedings of the Petroleum-
Hydrocarbons and Organic Chemicals in Ground
Water, Prevention, Detection, and Restoration
Conference and Exposition, Houston, TX.
November 4-6, 1987.
Johnson, R. L. and J. F. Pankow, 1987.
"Contaminant Transport in Fractured Porous
Rock; A Survey of Existing Ground-water
Contamination Sites and a Summary of Existing
Research Tools." Final Report, Cooperative
Research Agreement CR812553-010, R.S. Kerr
Environmental Research Laboratory, and U.S.
Environmental Protection Agency.
Kimmel, G. E. and O. C. Braids, 1980. "Leachate
Plumes in Groundwater from Babylon and Islip
Landfills, Long Island, New York." U.S. Geological
Survey, Professional Paper 1085.
Long, J. C. S., et al., 1982. "Porous Media
Equivalents of Discontinuous Fractures." Water
Resources Research, Vol. 18, No. 3, pp. 645-658.
MacFarland, D. S., et al., 1983. "Migration of Con-
taminants in Groundwater at a Landfill: A Case
Study; 1. Groundwater Flow and Plume
Delineation." Journal of Hydrology, Vol. 63, pp. 1-
29.
Matthess, G. and A. Pekdeger, 1981. "Concepts of a
Survival and Transport Model of Pathogenic
Bacteria and Viruses in Groundwater." Science of
the Total Environment, Vol. 21, pp. 149-159.
McDowell-Boyer, L.M., J.R. Hunt, and N. Sitar,
1986. "Particle Transport Through Porous Media."
Water Resources Research, Vol. 22, No. 13, pp.
1901-1921.
Mercado, A., 1967. "The Spreading Pattern of
Injected Water in a Permeability Stratified
Aquifer." Proceedings of the Symposium on
Artificial Recharge and Management of Aquifers,
International Association of Scientific Hydrology,
Vol. 72, pp. 23-36.
Molz, F. J., et al., 1982. "Performance, Analysis, and
Simulation of a Two-well Tracer Test at the Mobile
Site." Water Resources Research, Vol. 22, pp. 1031-
1037.
21
-------
Neretnieks, I., 1985. "Transport in Fractured Rocks."
In: Proceedings, Memoirs of the 17th International
Congress of the International Association of
Hydrotogists, Vol. 17, International Association of
Hydrologists, Tucson, AZ, 1985, pp. 301-308.
Neuman, S. P., et al., 1985. "Statistical Analysis of
Hydraulic Test Data From Fractured Crystalline
Rock Near Oracle, Arizona." In: Proceedings,
Memoirs of the 17th International Congress of the
International Association of Hydrologists, Vol. 17,
International Association of Hydrologists, Tucson,
AZ, pp. 289-300.
Neuzil, C. E., 1986. "Groundwater Flow in Low-
permeability Environments." Water Resources
Research, Vol. 22, pp. 1163-1195.
Novakowski, et al., 1985. "A Field Example of
Measuring Hydrodynamic Dispersion in a Single
Fracture." Water Resources Research, Vol. 21, pp.
1165-1174.
Pankow, J. F., et al., 1986. "An Evaluation of
Contaminant Migration Patterns at Two Waste
Disposal Sites on Fractured Porous Media in Terms
of the Equivalent Porous Medium (EPM) Model."
Journal of Contaminant Hydrology, Vol. 1, pp. 65-
76.
Perkins, T. K. and O. C. Johnston, 1963. "A Review
of Diffusion and Dispersion in Porous Media."
Journal of the Society of Petroleum Engineers, Vol.
17, pp. 70-83.
Pickens, J. F. and G. E. Grisak, 1981. "Scale-
dependent Dispersion in a Stratified Aquifer."
Water Resources Research, Vol. 17, pp. 1191-1212.
Pruess, K. and T. N. Narasimhan, 1985. "A Practical
Method for Modeling Fluid and Heat Flow in
Fractured Porous Media." Society of Petroleum
Engineers Journal, Vol. 25, No. 1, pp. 14-26.
Robin, M. J. L., R. W. Gillham, and D. W. Oscarson,
1987. "Diffusion of Strontium and Chloride in
Compacted Clay-based Materials." Soil Science
Society of America Journal, Vol. 51, pp. 1102-1108.
Sudicky, E. A., 1986. "A Natural Gradient
Experiment in a Sand Aquifer: Spatial Variability
of Hydraulic Conductivity and Its Role in
the Dispersion Process." Water Resources Research,
Vol. 22, No. 13, pp. 2069-2082.
Sudicky, E. A., R. W. Gillham, and E. O. Frind, 1985.
"Experimental Investigation of Solute Transport in
Stratified Porous Media: 1. The Nonreactive Case."
Water Resources Research, Vol. 21, No. 7, pp. 1035-
1041.
Tsang, Y. W. and C. F. Tsang, 1987. "Channel Model
of Flow Through Fractured Media." Water
Resources Research, Vol. 23, No. 3, pp. 467-479.
Van Genuchten, M. Th. and W. A. Jury, 1987.
"Progress in Unsaturated Flow and Transport
Modeling." Reviews of Geophysics, Vol. 25, No. 2,
pp. 135-140.
White, R. E., 1985. "The Influence of Macropores on
the Transport of Dissolved Matter Through Soil."
Advances in Soil Science, Vol. 3, pp. 95-120.
Witherspoon, P. A., et al., 1980. "Validity of Cubic
Law for Fluid Flow in a Deformable Rock
Fracture." Water Resources Research, Vol. 16, No.
6, pp. 1016-1024.
Wood, W. W. and G. G. Ehrlich, 1978. "Use of Baker's
Yeast to Trace Microbial Movement in Ground
Water." Groundwater, Vol. 16, No. 6, pp. 398-403.
Yates, M. V., et al., 1987. "Modeling Virus Survival
and Transport in the Subsurface." Journal of
Contaminant Hydrology, Vol. 1, pp. 329-345.
22
-------
CHAPTER 3
PHYSICAL PROCESSES CONTROLLING THE TRANSPORT OF
NON-AQUEOUS PHASE LIQUIDS IN THE SUBSURFACE
Carl D. Palmer and Richard L Johnson
Introduction
Liquids that do not readily dissolve in water and can
exist as a separate fluid phase are known as non-
aqueous phase liquids (NAPLs). Generally, NAPLs
are subdivided into two classes: those that are lighter
than water (LNAPLs); and those with a density
greater than water (DNAPLs). Most LNAPLs are
hydrocarbon fuels such as gasoline, heating oil,
kerosene, jet fuel, and aviation gas. Most DNAPLs
are chlorinated hydrocarbons such as 1,1,1-
trichloroethane, carbon tetrachloride, chlorophenols,
chlorobenzenes, tetrachloroethylene, and PCBs.
Concern about NAPLs exists because of their
persistence in the subsurface and their ability to
contaminate large volumes of water. For example, 7
L (10 kg) of trichloroethylene (TCE) can contaminate
108 L of ground water at 100 ppb (Feenstra and
Cherry, 1987). NAPLs are ubiquitous throughout
North America and have been identified at 4 out of 5
hazardous waste sites in the United States (Plumb
and Pitchford, 1985). Greater understanding of the
transport and dissolution of NAPLs is necessary to
implement cost-effective techniques for the cleanup
of these contaminants.
Transport and Dissolution of NAPLs
As NAPLs move through geologic media, they can
displace water and air. Because water is the wetting
phase relative to both air and NAPLs, it tends to line
the edges of the pores and cover the sand grains. The
NAPL is the non-wetting phase and tends to move
through the central portions of the pores. Neither the
water nor the NAPL phase occupies the entire pore.
Because of this, the permeability of the medium with
respect to these fluids is different than when the pore
space is entirely occupied by a given phase. This
reduction in permeability depends on the medium
and often is described in terms of relative
permeability, krj, for phase i, which is defined as:
k . = k.(S.)/k .
ri i i si
(D
where Si is the fraction of pore space occupied by
phase i, kj{Sj) is the permeability of the medium to
phase i at saturation S;, and ksi is the permeability of
the medium at complete saturation with phase i.
Thus, the relative permeability varies from 1.0 at
100 percent saturation to 0.0 at 0 percent saturation.
A plot of relative permeability versus water
saturation for a hypothetical medium (Figure 23)
reveals some important features about multiphase
flow. At 100 percent water saturation, the relative
permeability of the water and the NAPL are 1.0 and
0.0, respectively. As the fraction of the pore space
occupied by the NAPL (Sn) increases, a corre-
sponding decrease occurs in the fraction of water
within the pore space (Sw). As Sw decreases, the
relative permeability with respect to the water phase
decreases to zero. Zero relative permeability is not
obtained at zero Sw, but at the irreducible water
saturation (Srw). At this water saturation, the water
phase is effectively immobile and there is no
significant flow of water. These concepts are similar
to those discussed for unsaturated flow in Chapter 2.
The relative permeability of the NAPL behaves in a
similar manner. At 100 percent NAPL saturation,
the relative permeability for the NAPL is equal to
1.0, but as the NAPL saturation decreases, so does
the relative permeability. At the residual NAPL
saturation (Srn), the relative permeability for the
23
-------
100%
NAPL SATURATION
Irreducible
- Water
Saturation
Srw
WATER SATURATION
100%
Figure 23.
Relative permeability as a function of satura-
tion.
NAPL is effectively zero and the NAPL is considered
to be immobile.
These immobile fractions of NAPL cause great
concern because they cannot be easily removed from
the pores except by simple dissolution by flowing
ground water. An example of this could be a cubic
meter of soil with a 35 percent porosity and
containing TCE at a 20 percent residual saturation.
This situation implies that there is 0.07 m3 or 103 kg
of TCE within the soil. If the solubility of the TCE is
1,100 mg/L and if ground water flows through the
soil at a rate of 1.7 cm/day, it would take 15.4 years
for the TCE to be removed by dissolution. If the
contaminated aquifer is twice as long (2 m), 30.8
years are necessary. If a NAPL with a lower
solubility than TCE is spilled, the rate of mass
removal is lower, requiring even more time for
dissolution. Thus, NAPLs that enter the subsurface
can remain for decades and can contaminate large
volumes of ground water. Some of the key aspects of
NAPL dissolution are considered in more detail in
Chapter 6.
Understanding how an NAPL moves within a porous
aquifer can be useful. The movement of petroleum
products in the subsurface is described in detail by
Schwille (1967), and the movement of DNAPLs is
examined closely by Schwille (1988), Feenstra and
Cherry (1987), Kueper and McWorter (1988), and
Anderson (1988).
Light Non-Aqueous Phase Liquids
(LNAPLs)
As a spilled LNAPL enters the unsaturated zone, it
flows through the central portion of the unsaturated
pores. If the amount of product released is small, the
product flows until residual saturation is reached
(Figure 24A). Therefore, a three-phase system
consisting of water, product, and air is created within
the vadose zone. Infiltrating water dissolves the
components within the LNAPL (e.g.',' benzene,
toluene, and xylene) and Carries them to the water
table. These dissolved constituents then form a
contaminant plume emanating from the area of the
residual product. Many of the components commonly
found in LNAPLs are volatile and, as a consequence,
can partition into the soil air and be transported by
molecular diffusion to other portions of the aquifer.
As these vapors diffuse into adjacent soil areas, they
partition back into the water phase and spread
contamination over a wider area. If the surface is not
covered with an impermeable material, these vapors
diffuse, across the surface boundary and into the
atmosphere. However, if'a relatively impermeable
boundary covers the area, np,,mass transfer occurs
with the atmosphere and the'Concentrations of
contaminants in the soil atmosphere may build up to
equilibrium concentrations;
If large volumes of product are spilled (Figure 24B),
the product flows through the pore space to the top of
the capillary fringe. The dissolved components of the
infiltrating product precede the product and may
change the wetting properties of the water, causing a
reduction in the residual water, Content and collapse
of the capillary fringe. ;
The LNAPL product is lighter than water and tends
to float on top of the capillary fringe^ AS the he,ad
created by the infiltrating /prodiict increases, the
water table is depressed and the product begins to
accumulate in the depression. If the source of the
spilled product is then turned off, the LNAPL within
the vadose zone continues to flow under the influence
of gravity until reaching residual saturation. As this
drained product continues to recharge the product
pool, it spreads laterally on the top of the capillary
fringe (Figure 24C). The draining, of the upper
portions of the vadose zone also reduces the total
head at the interface between the product and the
ground water, causing the water table to rebound
slightly. The rebounding water can only displace a
portion of the product because the latter remains at
residual saturation. Ground water passing through
this area of residual saturation dissolves the com-
ponents within the residual product, creating a
contaminant plume. Water infiltrating from tne
surface also can dissolve the residual product and
vapors within the vadose zone, thereby contributing
to the overall contaminant load to the aquifer.
If the water table drops because , of 'seasonal
variations or pumping, the pool of product also drops.
If the water table rises again, part of the product is
pushed upward, but a portion remains at residual
saturation below the new water table,. Thus,
variations in the water table can spread'the product
24
-------
PRODUCT SOURCE
ttttt
GROUNDWATER
GROUNDWATER WATERTABLE
PRODUCT SOURCE
Httt
OROUNDWATER
GROUNDWATER
WATERTABLE
Figure 24. Movement of LNAPLS into the subsurface:
(a) distribution of LNAPL after small volume
has been spilled; (b) depression of the
capillary fringe and water table; (c)
rebounding of the water-table as LNAPL
drains from overlying pore space.
over a greater thickness of the aquifer, causing
increased volumes of soil to be contaminated. Clean-
up methods for LNAPLs in the subsurface should
take this principle into account and avoid moving the
product into uncontaminated areas where more
product can be held at residual saturation.
Dense Non-Aqueous Phase Liquids
(DNAPLs)
DNAPLs can have great mobility in the subsurface
as a result of their relatively low solubility, high
density, and low viscosity. The sparingly soluble
DNAPLs do not readily mix with water and therefore
remain as separate phases. The relatively high
density of these liquids provides a driving force that
can carry product deep into aquifers. The
combination of this high density and low viscosity is
particularly important with regard to the transport
of DNAPLs in the subsurface. When a high density,
low viscosity fluid (DNAPL) displaces a lower
density, higher viscosity fluid (water), the flow is
"unstable" and viscous fingering occurs (Saffman
and Taylor, 1958; Homsy, 1987; Chouke, et al., 1959;
Kueper and Frind, 1988).
During a spill (Figure 25A), DNAPL flows through
the unsaturated zone under the influence of gravity
toward the water table. If only a small amount of
DNAPL is spilled, it flows until reaching residual
saturation in the vadose zone. If there is water
within the unsaturated zone, the DNAPL exhibits
viscous fingering during infiltration. No viscous
fingering is observed if the vadose zone is dry. The
DNAPL can partition into the vapor phase and these
dense vapors may sink to the capillary fringe.
Infiltrating water can dissolve the residual DNAPL
or the vapors and transport these contaminants to
the water table, creating a dissolved chemical plume
within the aquifer.
If a greater amount of DNAPL is spilled (Figure
25B), the DNAPL flows until it reaches the capillary
fringe and, once there, begins to penetrate into the
aquifer. However, to do this, the DNAPL must
displace the water by overcoming the capillary forces
between the water and the medium. The critical
height of DNAPL required to overcome these
capillary forces (zc) can be calculated from:
= 2y cos (9) (l/rt - 1/r )/ (Ap g)
(2)
where y is the interfacial tension between the water
and the DNAPL, 9 is the contact angle between the
fluid boundary and the solid surface, rt is the radius
of the pore throat, rp is the radius of the pore, Ap is
the difference in the density between the water and
the DNAPL, and g is the acceleration of gravity
(Villaume, et al., 1983). As an example, calculated
critical heights' required for perchloroethylene to
penetrate saturated porous media of different grain
size range from a few centimeters for coarse grains to
tens of meters for clays (Table 2). Thus, unfractured,
saturated clays and silts can be effective barriers to
the migration of DNAPLs, provided the critical
heights are not exceeded-
After penetrating the aquifer, the DNAPL continues
to move through the saturated zone until it reaches
residual saturation. The DNAPL then is dissolved by
ground water passing through the contaminated
area, resulting in a contaminant plume that can
extend over a great thickness of the aquifer^ If finer-
grained strata are contained within the aquifer, the
25
-------
Table 2. Critical Height for Perchloroethylene to
Penetrate Water Saturated Media (Anderson,
1988).
DNAPL SOURCE
tttttt
Material
Coarse Sand
Fine Sand
Silt
Clay
Diameter
1.0
0.1
0.01
0.001
Critical
Height (cm)1
13
130
1,300
13,000
'Calculated tor Ap » 0.62 g/cm^.
Y ™ 47.5 dynes/cm, cos 9 = 1.
infiltrating DNAPL accumulates on top of the
material, creating a pool. At the interface between
the ground water and the DNAPL pool, the solvent
dissolves into the water and spreads vertically by
molecular diffusion. As water flows by the DNAPL
pool, the concentration of the contaminants with the
ground water increases until saturation is achieved
or the downgradient edge of the pool is reached. The
relative density of pools and fingers of DNAPL
within the aquifer is important when controlling the
measured concentrations of dissolved contaminants
derived from DNAPLs. The existence of fingers and
pools of the DNAPL, rather than relatively
continuous distributions, in the subsurface accounts
for the observation that the concentration of many of
the DNAPL compounds in ground water are far
below their saturation limit (Anderson, 1988).
If even larger amounts of DNAPL are spilled (Figure
25C), the DNAPL can, in principle, penetrate to the
bottom of the aquifer, forming pools in depressions. If
the impermeable boundary is sloping, the DNAPL
flows down the dip of the boundary. This direction
can be upgradient from the original spill area if the
impermeable boundary slopes in that direction. The
DNAPL also can flow along bedrock troughs, which
may be oriented differently from the general
direction of ground-water flow. This flow along low
permeability boundaries can spread contamination
in directions that would not be predicted on the basis
ofhydraulics.
The transport of DNAPLs in physical models of
fractures illustrates the importance of fracture
aperture and roughness. Schwille (1988) found that
if fracture apertures are greater than 0.2 to 0.5 mm,
the DNAPL moves directly to the capillary fringe
where it spreads out. Eventually the DNAPL
penetrates the capillary fringe and is transported to
the bottom of the aquifer. There is very little residual
DNAPL for fractures larger than 0.2 mm. If the
fracture aperture is less than 0.2 mm and is smooth,
the DNAPL spreads out near the surface and a few
fingers migrate down to the capillary fringe. At the
capillary fringe, the DNAPL spreads out further and
a few relatively wide fingers penetrate below the
B
DNAPL SOURCE
tttttt
Figure 25. Movement of DNAPLs into the subsurface:
(a) distribution of DNAPL after small volume
has been spilled; (b) distribution of DNAPL
after moderate volume has been spilled; (c)
distribution of DNAPL after large volume has
been spilled (After Feenstra and Cherry,
1988).
capillary fringe. If the fracture is rough, there is a
great amount of fingering and the DNAPL
penetrates below the capillary fringe in small,
scattered fingers. Thus, for fractures with apertures
less than 0.2 mm, there can be a large volume of
DNAPL that remains at residual saturation in the
fractures both above and below the capillary fringe.
While similar behavior is expected to occur in
fractured rock, the statistical distribution of fracture
26
-------
aperture and fracture roughness may preclude the
use of such a simple categorization based on the
mean apertures.
References
Anderson, M. R., 1988. "The Dissolution and
Transport of Dense Non-aqueous Phase Liquids in
Saturated Porous Media." Ph.D Dissertation.
Department of Environmental Science and
Engineering, Oregon Graduate Center, Beaverton,
OR.
Chouke, R. L., P. van Meurs, and C. van der Poel,
1959. "The Instability of Slow, Immiscible, Viscous
Liquid-liquid Displacements in Permeable Media."
Transactions AIME, Vol. 216, pp. 188-194.
Feenstra, S. and J. A. Cherry, 1987. "Dense Organic
Solvents in Ground Water: An Introduction." In:
Dense Chlorinated Solvents in Ground Water,
Institute for Ground Water Research, University of
Waterloo, Waterloo, Ontario, Progress Report
0863985.
Homsy, G. M., 1987. "Viscous Fingering in Porous
Media." Annual Review of Fluid Mechanics, Vol.
19, pp. 271-311.
Kueper, B. H. and E. O. Frind, 1988. "An Overview
of Immiscible Fingering in Porous Media." Journal
of Contaminant Hydrology, Vol. 2, pp. 95-110.
Kueper, B. H. and D. B. McWorter, 1988. "Mechanics
and Mathematics of Immiscible Fluids in Porous
Media." In: Dense Chlorinated Solvents in Ground
Water, Institute for Ground Water Research, Uni-
versity of Waterloo, Waterloo, Ontario, Progress
Report 0863985.
Plumb, R. H., Jr. and A. M. Pitchford, 1985. "Volatile
Organic Scans: Implication for Ground-water
Monitoring." Proceedings Petroleum Hydrocarbons
and Organic Chemicals in Ground Water, National
Water Well Association, November 13-15, 1985,
Houston, TX, pp. 207-222.
Saffman, P. G. and G. Taylor, 1958. "The Penetration
of a Fluid into a Porous Medium or Hele-Shaw Cell
Containing a More Viscous Liquid." Proceedings of
the Royal Society London, A(245), pp. 312-329.
Schwille, F., 1967. "Petroleum Contamination of the
Subsoil - A Hydrological Problem." In: The Joint
Problems of the Oil and Water Industries, Peter
Hepple, Editor. Elsevier, Amsterdam, pp. 23-53.
Schwille, F., 1988. Dense Chlorinated Solvents in
Porous and Fractured Media. Translated by J.F.
Pankow. Lewis Publishers, Inc., Chelsea, MI.
Villaume, J. F., P. C. Lowe, and D. F. Unites, 1983.
"Recovery of Coal Gasification Wastes: An Innova-
tive Approach." Proceedings of the Third National
Symposium on Aquifer Restoration and Ground-
Water Monitoring. National Water Well Associ-
ation, Worthington, OH, pp. 434-445.
27
-------
-------
CHAPTER 4
DETERMINATION OF PHYSICAL TRANSPORT PARAMETERS
Carl D. Palmer and Richard L Johnson
Introduction
The models used for the simulation and prediction of
contaminant transport in the subsurface are only as
good as the data input used in those models. The
range of important variables for contaminant
transport are tabulated (Mercer, et al., 1982), but
ultimately site-specific values must be obtained.
Hydrogeological parameters such as hydraulic
conductivity, porosity, bulk density, ground-water
flux, and dispersion are important for modeling;
methods for determining these parameters are
discussed below.
Both laboratory and field methods are useful for
determining hydrogeological parameters. Although
parameters measured in the laboratory are
applicable in small-scale situations, they may not be
representative of the bulk properties of the
formation. Still, if many such measurements are
taken, not only can the average, larger scale
properties of the formation be estimated, but other
important statistical properties also can be calcu-
lated. For example, Sudicky (1986) applied such
techniques to determine the mean and variance of
the saturated hydraulic conductivity of the Borden
aquifer in order to calculate the macroscopic scale
dispersivity.
Field methods represent an increase in the scale of
measurement relative to most laboratory methods.
This increase does not mean that such methods are
inherently better than laboratory methods, but
simply that the field-measured variables represent
average properties over a larger volume. The key
advantage of field methods is the potential for
measuring less disturbed materials, thereby giving
more accurate representations of the relevant
parameters. The disadvantage to using such methods
is that during the data analysis, ideal models are
applied to non-ideal media. To reduce this
discrepancy, field and laboratory methods should be
developed or modified to complement one another.
Hydraulic Conductivity
Hydraulic conductivity can be measured both in the
laboratory and the field. Laboratory methods include
estimation from grain-size analysis, permeameter
tests, and soils-engineering tests. Slug tests, aquifer
tests, and flow net analyses are field methods that
provide increasing scales of measurement.
Laboratory Methods
Hydraulic conductivity, K, can be estimated from
grain-size distribution curves using either the Hazen
method (Freeze and Cherry, 1979) or the Masch and
Denny method (1966). Sieve, pipet, hydrometer, and
settling tube methods can be used to analyze grain
size, and light scattering techniques have been
developed to look at micron- and submicron-sized
particles. Sample preparation and methods for
particle-size analysis are described in detail by Gee
and Bauder (1986). In addition to its use for
estimating hydraulic conductivity, grain-size
analysis can be used to properly size filter packs and
screens for monitoring, extraction, and injection
wells (Driscoll, 1986).
Saturated hydraulic conductivity is measured with
either constant-head or falling-head permeameter
tests. Constant-head tests are useful for measuring
hydraulic conductivity in the range of 100 to 10-5
29
-------
cm/s. Falling-head tests work best over a range of
10-3 to 10'7 cm/s. Undisturbed samples used in
permeameter tests offer the best results. If the soil
structure is disturbed, errors can be minimized by
repacking the soil to its original bulk density.
Soils-engineering tests in consolidometers and
triaxial cells provide coefficients of compressibility
and consolidation of soils. These properties are
related to the hydrogeologic parameters of hydraulic
conductivity, K, and specific storage, Ss (Jorgensen,
1980). Parameters K and Ss for fine-grained
materials have been successfully calculated from the
coefficients of compressibility and consolidation
(Desaulniers, et al., 1981; Paul, 1987). In fractured
tills, these calculated values of K and Ss represent
values for the matrix.
Unsaturated hydraulic conductivity can be meas-
ured using steady-state head control methods,
steady-state flux control methods, non-steady state
methods, or sorptivity methods. Details of methods
for both saturated and unsaturated media are
provided by Klute and Dirksen (1986). Empirical
models for predicting unsaturated hydraulic
conductivity as a function of water content or
pressure head are described by Mualem (1986).
Field Methods
Slug tests are the most common method for obtaining
hydraulic conductivity in the field. These tests are
conducted by instantaneously changing the
hydraulic head within a well and measuring its
return to the static level. Different types of slug tests
include: (1) falling head; (2) rising head; (3) bail; and
(4) pressure/packer.
In a falling-head test, the water level is instan-
taneously increased by adding a slug of water, or,
preferably, some displacing volume of material such
as a metal rod. The head is then measured as it falls
back to its static level. In a rising head test, a volume
(e.g., a metal rod) is removed from the well, causing
the water level to instantaneously drop. The rising
hydraulic head is then measured over a time
interval.
A bail test is a type of rising head test where the
water is removed using a bailer. This technique is
suitable in low-permeability material where the
time needed to remove the water is short relative to
the overall time needed for recovery in the well. In a
pressure/packer test, an interval of the well is
isolated by packers and a pressure pulse is applied to
that area. The decay of the pressure pulse can be
measured with a pressure transducer and the data
interpreted in the same manner as the other test
methods. In principle, these data are used to
calculate the hydraulic conductivity and the specific
storage of the geologic material.
Several techniques used for analyzing slug test data
are described by Hvorslev (1951), Bouwer and Rice
(1976), Cooper, et al. (1967), and Nguyen and Finder
(1984). Both the Hvorslev (1951) and the Bouwer and
Rice (1976) methods are based on steady-state flow
equations. If the specific storage of the medium is
small, these techniques can provide a good estimate
of the hydraulic conductivity. In theory, the relative
recovery in the well should plot as a linear function
of time for both of these methods. The slope of this
recovery-time curve is used to calculate hydraulic
conductivity. If the formation has a non-zero specific
storage, a curve rather than a straight line is
obtained and application of steady-state equations
results in over-estimation of the hydraulic conduc-
tivity.
The Cooper, et al. (1967) and the Nguyen and Pinder
(1984) methods are based on transient ground-water
flow equations and, therefore, better represent the
physical conditions within the soil. The Cooper, et al.
(1967) method assumes that the well fully penetrates
the aquifer while the Nguyen and Pinder (1984)
method accounts for partial penetration of the well.
The Cooper, et al. (1967) method is a curve-fitting
technique and is not very sensitive to the value of the
specific storage. The Nguyen and Pinder (1984)
model is the most general and, in principle, should be
the best technique for interpreting slug test data.
Instead of being a curve-fitting technique, the model
uses the slopes of two different plots to calculate the
specific storage and the hydraulic conductivity.
Many factors contribute to errors in the calculation
of hydraulic conductivity from slug test data (Table
3). Most of these factors affect the estimation of
hydraulic conductivity by a factor of 2 or 3. An
important exception is when a low-permeability skin
forms at the well-bore interface. Under these
circumstances, order of magnitude errors can result
(Palmer and Paul, 1987; Faust and Mercer, 1984). In
addition, simulated recovery data from wells that
have low permeability skins of finite thickness at the
well-bore interface can produce straighter lines on a
Hvorslev plot than simulated data for ideal wells
without a skin. Because of this, a straight-line
Hvorslev plot is not necessarily a valid criterion for a
"good" slug test.
Aquifer tests can provide larger scale information
about hydraulic properties than laboratory methods
or slug tests, and can be used to determine hydraulic
conductivity, specific storage, leakage, aquitard
diffusivity, anisotropy, and the general location of
boundaries. Also, aquifer tests can be constant rate,
variable rate, or constant head. Many different types
30
-------
Table 3. Potential Sources of Error in Slug Tests (Palmer
and Paul, 1987)
Bridging of Seals
Leaky Joints
Formation of Low Permeability Skins
Entrapped Air
Presence of Fractures
Stress Release Around Borehole
Partial Penetration of Well
Anisotropy of Formation
Varying Regional Potentiometnc Surface
Boundary Conditions
Sand Pack Effects
Uncertainty in Initial Head
Radius of Influence of Test
Thermal Expansion
of methods are used to analyze aquifer test data and
the choice of method depends on the conditions under
which the data were collected (steady-state flow or
non^steady-state flow) and the type of aquifer
(confined, unconfined, or semi-confined). Methods for
analyzing aquifer test data can be found in
Kruseman and de Ridder (1970), Walton (1962, 1970,
and 1984), Lohman (1972), and Hantush (1964).
The classic test method of Theis (1935) can be used
for confined aquifers. However, this method is
limited to fully penetrating wells in isotropic media.
If the wells partially penetrate the aquifer, a variety
of methods by Hantush (1964) can be used. These
methods also can be applied to leaky aquifers. If
monitoring wells are installed in the aquitard, then
the ratio method (Neuman and Witherspoon, 1972) is
useful for obtaining the hydraulic diffusivity of the
aquitard.
Investigators have long debated the nature of
"delayed yield" observed in aquifer tests conducted
in unconfined aquifers (Neuman, 1972 and 1979).
The methodology presented by Neuman (1975) has a
stronger physical basis than other methods and
accounts for anisotropy and partial penetration.
Neuman's (1975) method is highly recommended for
use in unconfined aquifers.
Although the methods described above for analyzing
aquifer test data were derived for porous media, they
also can be applied to aquifer tests conducted in
fractured rock if the aquifer behaves like a single
porosity medium. However, such fractured media are
often highly anisotropic, and the methods by Weeks
(1969) or by Way and McKee (1982) may be more
appropriate in these situations.
If the test is conducted in a fractured, porous aquifer,
then the double porosity methods by Barenblatt, et
al. (1960) or by Boulton and Streltsova (1977) may be
applied. Determining whether an aquifer is a single
or double porosity system can be accomplished using
a simple method (Gringarten, 1984), where a plot of
ds/3 [ln(t)l versus t is made on log-log paper (Figure
26). In a single porosity aquifer, the plot should
increase to a maxima, then decrease and become
constant. In a double porosity aquifer, the plot should
go through a maxima, and then decrease and pass
through a minima before finally becoming constant.
SINGLE POROSITY
DOUBLE POROSITY
LOG (t)
Figure 26. Differentiating double porosity media from
single porosity media (after Gringarten,
1984).
Flow-net analysis can provide a larger scale estimate
of hydraulic conductivity than aquifer tests. While
regional scale values may not be directly applicable
to a specific site, they may prove useful in regional
scale models to define local boundary conditions.
Methods of flow-net analysis are described in Freeze
and Cherry (1979) and Cedergren (1967).
Unsaturated hydraulic conductivity can be
measured in the field using either steady-state or
non-steady-state flux methods. Such techniques are
described in detail by Green, et al. (1986).
Bulk Density, Porosity, and Volumetric
Water Content
Bulk density, porosity, and volumetric water content
also are required for transport models. Porosity is
necessary to estimate the ground-water velocity from
the Darcy flux while both the porosity and the bulk
density are needed to calculate a retardation factor
from a partition coefficient (see Chapter 5). The
volumetric water content is required to estimate
water and contaminant movement through the
vadose zone. Laboratory techniques exist for
measuring these parameters (Danielson and
Sutherland, 1986; Blake and Hartage, 1986;
Gardner, 1986). Field methods used to estimate these
parameters include neutron logs for porosity and
water content, gamma-gamma logs for bulk density,
and electrical capacitance for water content. In
31
-------
addition, the porosity can be calculated from the
average linear velocity, v, and the ground-water
flux, q, by q/v.
Ground-Water Flux and Average Linear
Velocity
Potentiometric Surface Data
The most common method for estimating the
magnitude and direction of ground-water flux is with
potentiometric surface data and hydraulic conduc-
tivity. The direction of ground-water flow in isotropic
media is downgradient orthogonal to the equipo-
tential lines. The flux rate, q, can be calculated
directly from Darcy's equation and the product of
hydraulic conductivity and gradient. If the porosity,
0t, is known, the average linear velocity, v, is
calculated as the ratio, q/0t. If the aquifer is aniso-
tropic, the flow lines and the equipotentials are not
orthogonal and the direction of flow must be
estimated from flow nets constructed in the
transformed section or through use of the inverse
hydraulic-conductivity ellipse (e.g., Freeze and
Cherry, 1979).
Borehole Dilution
Borehole dilution is a relatively simple technique to
determine the magnitude and, in principle, the
direction of ground-water flow. In a borehole dilution
test, an interval within a well is isolated with
packers and a tracer is injected and continuously
mixed. During the test, ground water enters the well
bore and dilutes the tracer. The rate at which the
tracer is diluted within the well bore is a measure of
the rate of ground-water flow. The relationship
between the concentration and ground-water flux is:
In
(C - C')
CC0- C')
A0q
w " - V
(1)
where C' is the background concentration, C0 is the
concentration in the injected slug, A is the cross-
sectional area of the borehole, W is the volume in the
isolated section of the borehole, and q is the ground-
water flux. The P parameter accounts for the con-
vergence of flow lines on the open borehole and is
often called the borehole factor. This convergence of
flow lines results in a greater flux through the
borehole than through the aquifer. For an ideally
installed well with no sandpack and with the screen
permeability much greater than the formation, p has
a value of 2.0. Methods for estimating 0 are discussed
by Drost, et al. (1968) and by Halvely, et al. (1966).
According to Equation 1, a plot of the logarithm of
relative concentration versus time should produce a
straight line with a slope of - A{5q/W. Thus, if A, 0,
and W are known, the ground-water flux, q, can be
calculated. Experiments conducted in a sandbox by
McLinn and Palmer (1988) show that the logarithm
of relative concentration versus time is linear and
that the flux rates calculated from the borehole
dilution test agree with the measured flux values
(Figure 27).
q(calc.) = 5.9 cm/s
q(meas.) = 5.7 cm/s
100
200
300
400
TIME (MINUTES)
Figure 27.
Results of borehole dilution test illustrating
linear relationship between the logarithm of
relative concentration and time.
There are several different types of borehole dilution
devices. For example, early investigators used
radioisotopes with scintillation counters (Halvely, et
al., 1966), as well as specific ion electrodes (Grisak,
et al., 1977) and specific conductance electrodes
(Bellanger, 1985). Currently, a thermal device is
commercially available (Kerfoot and Skinner, 1980;
and Kerfoot, 1982), and an electrical resistivity
device has been developed (McLinn and Palmer,
1988). One key advantage of borehole dilution is its
potential to profile the distribution of velocities
along the length of the well screen. Under appro-
priate conditions, therefore, the device can be used to
discern velocity variations within the aquifer and to
describe aquifer heterogeneity.
Seepage Meters
Seepage meters can be an extremely useful tool for
measuring fluid flux at the ground-wate'r/surface-
water interface (Lee, 1977; Lee and Cherry, 1978).
32
-------
The seepage meter consists of an inverted section of a
55-gallon drum with a hole in the top that is covered
with a plastic bag (Figure 28). After a measured
period of time (At), the bag is removed and the
volume of water collected, V, is measured. The flux of
ground water through the sediments and into the
surface water is simply V/(AAt) where A is the area
of the drum. If a mini-piezometer is used in con-
junction with the seepage meter, the hydraulic
gradient between the sediments and the surface
water can be calculated. If the hydraulic gradient
and the flux rate are known, a simple calculation
using Darcy's equation yields the hydraulic conduc-
tivity of the sediments. The seepage meter works
best in materials that are not easily compressed; if
the sediments are compressed during installation,
the hydraulic conductivity will be underestimated.
Mini-piezometer
Plastic Bag
Figure 28. Seepage meter.
Dispersion Coefficient and Aquifer
Heterogeneity
Tracer Tests
Tracer tests have been used to study physical,
chemical, and biological processes in the subsurface.
Physical parameters such as dispersion, average
linear velocity, porosity, and variation in hydraulic
conductivity have been obtained from tracer tests.
The tracer tests by Sudicky, et al. (1983) and Button
and Barker (1985) show that ground-water velocity
can vary substantially over small scales. In their
experiments, small slugs of contaminants broke into
discrete parts that moved at different velocities.
Also, Sudicky, et al. (1983) and Freyberg (1986) used
tracer tests to describe how the apparent dispersivity
value increases with the scale of the problem.
Pickens and Grisak (1981) were able to demonstrate
that when the concentration breakthrough curves
obtained with point samplers are analyzed, the
dispersivity parameter is small (0.007 m). In
addition, Palmer and Nadon (1986) and Taylor, et al.
(1988) suggest that electrical resistivity can be
coupled with the execution of single-well injection
tracer tests to obtain information about variation in
hydraulic conductivity.
Tracer tests can be conducted in several ways.
Natural-gradient tracer tests are commonly used for
research purposes. In these tests, a slug of tracer is
injected into an aquifer and traced as it moves along
with the natural hydraulic gradient. Such tests were
used by Sudicky, et al. (1983), Sutton and Barker
(1985), and MacKay, et al. (1986).
Forced-gradient tracer-, experiments involve the
injection of a tracer over a prolonged period of time
and measurement of the arrival of the tracer at one
or more monitoring wells during the injection period.
Forced-gradient wells may be either single-well or
double-well tests. In single-well tests, only one
"active" well is used for injecting and withdrawing
the tracer. The fluid is injected into the aquifer and
monitored at one or more sampling wells at some
radius from the active well (Figure 29). Two-well
tests (Figure 30) use two active wells, one to inject
the tracer and the other for withdrawal. These types
of forced gradient tests were used by Pickens and
Grisak (1981), Molz, et al. (1986), Palmer and Nadon
(1986), and Taylor, et al. (1988).
Experiments by Molz, et al. (1986) demonstrate that
tracer tests can be applied as an engineering tool and
numerical models can be used to predict the
transport of solutes in the subsurface. Molz, et al.
(1986) initially conducted a single-well tracer test in
a confined aquifer to measure the variation in
hydraulic conductivity over the thickness of the
aquifer. These tests were on a scale of approximately
5.5 m. A two-well tracer test with travel distances of
38 to 90 m was also conducted. Using the hydraulic
conductivity distribution for the single-well test, the
concentration-versus-time curve of a pumping well
in a two-well test could be reasonably simulated
without recourse to model calibration (Figure 31).
Two models were used: (1) a three-dimensional
advection-dispersion model using small dispersivity
values; and (2) an advection model. The results of the
two models are indistinguishable. If the aquifer is
assumed to be homogeneous and the concentrations
33
-------
INJECTION
C|NJ(t)
UPPER
WITHDRAWAL
Q=00UT
e.
•K(z)
CONFINING LAYER
INJECTION-
-WITHDRAWAL
WELL
OBSERVATION
WELL —2 .
WITH
MULTILEVEL
SAMPLERS
-•-r
LOWER CONFINING LAYER
Figure 29. Vertical cross-section showing single-well
test geometry (Molz, et al., 1986).
penetrating radar, induced electrical polarization,
resistivity, magnetometer, reflection seismics, and
electromagnetic surveys. Borehole methods include
geothermetry, electrical methods, acoustic methods,
and nuclear logging techniques. Surface geophysical
techniques such as resistivity, conductivity, seismic
refraction, and VLF can be useful in identifying
lithology changes in the subsurface, depth to water
table, and depth to bedrock. Ground-penetrating
radar, electromagnetic methods (EM), and resistivity
are effective in locating buried drums and
containers. Borehole geophysical techniques are
used to estimate hydraulic conductivity (resistivity),
lithology (natural gamma), bulk density (gamma-
gamma), and porosity (neutron). Discussions of these
topics can be found in Keys and MacCary (1971),
Benson, et al. (1983), and Rehm, et al. (1985). As
described in the discussion of tracer tests, borehole
resistivity logs can be coupled with tracer tests to
obtain information about subsurface heterogeneity
(Taylor, et al., 1988; Palmer and Nadon, 1986).
Injection well
(source)
Withdrawal well
(sink)
Plan view
Multi-Level
Observation
Well
xxx^xxxyxl Uxxx^
If
h -!*
i
i
•
/•yxyyxy
\
~KMJ
J
/
sssssssA \s
I
r*
1
I r*
it
SA \SSSSSS////S
-jj-*
"{h*
j i
"ir*"
Vertical section in x-z plane
Figure 30. Two-well test geometry in a stratified aquifer
(Molz, et al., 1986).
in the pumping well are calculated using average
hydraulic properties, poor replication of the
experimental data is obtained (Figure 32). These
results emphasize the importance of heterogeneity
and advection to the transport of solute through
porous media.
Geophysical Techniques
While geophysical techniques are recognized as
useful tools in characterizing waste sites, a detailed
discussion of geophysical techniques is beyond the
scope of this chapter. Surface geophysical techniques
used include gravity, infrared imagery, ground-
Plume Detection
Ground-Water Monitoring
The most common method for detecting contaminant
plumes in the subsurface is by direct sampling of the
subsurface fluid. However, the common practice of
"plume chasing" may not be the most efficient
method for determining the extent of ground-water
contamination (Dowden and Johnson, 1988). Greater
use of hydrogeologic data can greatly improve the
design of ground-water monitoring systems and the
quality of information provided.
Although ground-water monitoring networks can
provide the greatest certainty about the extent of
contamination, there are important factors that
control the quality of this information, including the
amount of well purging done prior to sampling
(Barcelona and Helfrich, 1986), the method of
sampling (Stolzenburg and Nichols, 1985), and the
method of well construction and installation (Keely
and Boateng, 1987). Methods for ground-water
sampling are discussed in detail by Scalf, et al.
(1981), Ford, et al. (1984), and Barcelona, et al.
(1985).
Geophysical Techniques
Geophysical methods can be used to locate
contaminant plumes with high dissolved solids. For
example, surface resistivity methods have been
successful at some sites (e.g., Stellar and Roux, 1975;
Kelly, 1976; Rogers and Kean, 1980), and electro-
magnetic methods are useful for obtaining some
estimate of water quality changes in the subsurface
34
-------
^
\
0
—
Z
O
h-
o:
i-
z
UJ
u
z
o
u
25.0
22.5
20.0
1 7.5
1 5.0
1 2.5
1 0.0
7.5
5.0
2.5
n
_
-
-
-
-
.
-
-
~
,
0
—•—Experiment
Geotrans Calculation (3-D)
0 TWAM Calculation
(Advection)
•J 1 1 1 1 1 1 \ III'
Figure 31.
40 80 I 20 160 200 240 280 320 360 400 440 480 520 560 600 640 680 720 760
TIME (MRS)
Comparison of predicted and modeled concentrations in the pumping well of a two- well tracer test (Molz,
etal., 1986).
(Stewart, 1982; Slaine and Greenhouse, 1982;
Glaccum, et al., 1982; Ludwig, 1983). The borehole
counterparts to these surface geophysical methods
should work equally as well in determining vari-
ations in water quality over the thickness of the
aquifer. Detailed discussions of these methods and
their limitations can be found in Keys and MacCary
(1971), Benson, et al. (1983) and Rehm, et al. (1985).
Chemical Time-Series Sampling Tests
Chemical time-series sampling tests are conducted
by repeatedly sampling wells that are continuously
pumped (Keely, 1982). The concentration data can
then be plotted as a function of time, volume
removed, or an equivalent radius from the well bore.
The particular shape of the curve obtained from such
a test varies according to the distribution of
contaminants in the subsurface. Some suggested
interpretations are provided in Figure 33. Data from
the time-series test have been used to determine the
proper amount of well purging that should be
performed before collecting water samples at the
site, and to identify the source of a contaminant
entering a water-supply well (Keely and Wolf, 1983).
While there is no unique interpretation of the
resultant curve, its general shape can provide
information that must be reconciled with hypotheses
concerning the distribution of contaminants near the
well bore.
References
Barcelona, M. J., et al., 1985. "Practical Guide for
Ground-water Sampling." U.S. EPA Report,
EPA/600/2-85/104. Robert S. Kerr Environmental
Research Laboratory, Office of Research and
Development, U.S. Environmental Protection
Agency, Ada, OK.
Barcelona, M. J. and J. A. Helfrich, 1986. "Well
Construction and Purging Effects on Ground-water
Samples." Environmental Science and Technology,
Vol. 20, pp. 1179-1184.
Barenblatt, G. I., I. P. Zheltor, and I. N. Kochina,
1960. "Basic Concepts in the Theory of Seepage of
Homogeneous Liquids in Fissured Rocks."
Prikladnaya Matem, Mekh (USSR), Vol. 24, No. 5,
pp. 852-864.
Bellanger, D., 1985. "Ground-water Velocity
Measurements Using An Electrical Conductance-
Borehole Dilution Device." Unpublished Master's
35
-------
<
r-
•z.
UJ
o
z
o
o
45
40
35
3O
25
I 5
I O
8
o o
o
0° °
o
O 4O 8O ISO I6O ZOO 240 2BO 32O 360 400 44O 480 320 560 600 640 680 720 760
TIME (hrs.)
Figure 32, Comparison of predicted and modeled concentrations in the pumping well of two-well tracer test
assuming a homogeneous aquifer (Molz, et al., 1986).
Thesis, Department of Earth Sciences, University
of Waterloo, Waterloo, Ontario.
Benson, R. C., R. A. Glaccum, and M. R. Noel, 1983.
"Geophysical Techniques for Sensing Buried
Wastes and Waste Migration." Contract Report No.
68-03-3050, Environmental Monitoring Systems
Laboratory, Office of Research and Development,
U.S. Environmental Protection Agency, Las Vegas,
NV.
Blake, G. R. and K. H. Hartage, 1986. "Bulk
Density." In: Method of Soil Analysis, Part I,
Physical and Mineralogical Methods, A. Klute,
Editor. Agronomy, No. 9, Part 1, American Society
of Agronomy, Inc., Soil Science Society of America,
Inc., Madison, WI, pp. 363-375.
Boulton, N.S. and T.D. Streltsova, 1977. "Unsteady
Flow to a Pumped Well in a Fissured Water
Bearing Formation." Journal of Hydrology, Vol.
35, pp. 257- 270.
Bouwer, H. and R.C. Rice, 1976. "A Slug Test for
Hydraulic Conductivity of Unconfined Aquifers
with Completely or Partially Penetrating Wells."
Water Resources Research, Vol. 12, No. 3, pp. 423-
428.
Cedergren, H., 1967. Seepage, Drainage, and Flow
Nets. John Wiley & Sons, New York.
Cooper, H. H., J. D. Bredehoeft and I. S.
Papadopulos, 1967. "Response of a Finite-diameter
Well to an Instantaneous Charge of Water." Water
Resources Research, Vol. 3, No. 3, pp. 263-269.
36
-------
o
<
cc
UJ
o
o
o
Figure 33A.
O
I
O
o
Figure 33B.
TIME OR VOLUME PUMPED
Hypothetical curve obtained from chemical
times-series sampling test. (The results may
be interpreted as data from a monitoring
well placed at the edge of a plume where
the concentrations are reduced due to
increasing amounts of uncontaminated
water entering the borehole.) (After Keely,
1982).
B
CC
I
o
I
TIME OR VOLUME PUMPED
Hypothetical curve obtained from chemical
times-series sampling test. (The results
may be interpreted as data from a
monitoring well that may have been
contaminated during drilling or installation.
The concentrations rapidly decrease as
uncontaminated water from the aquifer
enters the well bore.) (After Keely, 1982).
TIME OR VOLUME PUMPED
Figure 33C. Hypothetical curve obtained from chemical
times-series sampling test. (The results may
be interpreted as data from a monitoring
well in which chemical reactions have
reduced concentrations in the stagnant
water column. When pumping is started,
contaminated ground water from the aquifer
enters the well and contaminant
"concentrations rapidly increase.) (After
Keely, 1982).
Danie'lson, R.E. and P.L. Sutherland, 1986.
"Porosity." In: Method of Soil Analysis, Part I,
Physical and Mineralogical Methods, A. Klute,
Editor. Agronomy, No. 9, Part 1, American Society
of Agronomy, Inc., Soil Science Society of America,
Inc., Madison, WI, pp. 443-461.
Desaulniers, D. E., J. A. Cherry, and P. Fritz, 1981.
"Origin, Age, and Movement of Pore Water in
Argillaceous Quaternary Deposits at Four Sites in
Southwestern Ontario." Journal of Hydrology, Vol.
50, pp. 231-257.
Dowden, J. and L. Johnson, 1988. "Cost-benefit
Analysis of Alternative Remedial Investigation
Methodologies: A Case Study." Proceedings of the
Second National Outdoor Conference on Aquifer
Restoration, Ground-water Monitoring, and Geo-
physical Methods, Las Vegas, NV, May 23-26,
1988, pp. 1495-1507.
Driscoll, F.G., 1986. Groandwater and Wells, 2nd
Edition. Johnson Division, St. Paul, MN.
Drost, W. D., et al., 1968. "Point Dilution Methods of
Investigating Ground Water Flow by Means of
Radioisotopes." Water Resources Research, Vol. 4,
No. I, pp. 125-146.
Faust, C. R. and J. W. Mercer, 1984. "Evaluation of
Slug Tests in Wells Containing a Finite Thickness
Skin." Water Resources Research, Vol. 20, No. 4,
pp. 504-506.
Ford, P. J., P. J. Turina, and D. E. Seely, 1984.
"Characterization of Hazardous Waste Sites — A
Methods Manual, Volume II, Available Sampling
Methods." U.S. EPA Report, EPA/600/4-84/076.
Robert S. Kerr Environmental Research Labora-
tory, Office of Research and Development, U.S.
Environmental Protection Agency, Ada, OK.
Freeze, R. A. and J. A. Cherry, 1979. Groundwater.
Prentice-Hall, Inc., Englewood Cliffs, NJ.
Freyberg, D. L., 1986. "A Natural Gradient
Experiment on Solute Transport in a Sand Aquifer.
2. Spatial Moments and the Advection and Dis-
persion of Nonreactive Tracers." Water Resources
Research, Vol. 22, pp. 2031-2046.
Gardner, W. H., 1986. "Water Content." In: Methods
of Soil Analysis, Part /, Physical and Mineralogical
Methods, A. Klute, Editor. Agronomy, No. 9, Part
1, American Society of Agronomy, Inc., Soil Science
Society of America, Inc., Madison, WI, pp. 493-544.
Gee, G. W. and J. W. Bauder, 1986. "Particle-size
Analysis." In: Methods of Soil Analysis, Part I,
Physical and Mineralogical Methods, A. Klute,
Editor. Agronomy, N'o. 9, Part 1, American Society
of Agronomy, Inc., Soil Science Society of America,
Inc., Madison, WI, pp. 383-411.
Glaccum, R. A., R. C. Benson, and M.R. Noel, 198?
"Improving Accuracy and Cost-effectiveness of
37
-------
Hazardous Waste Site Investigations." Ground-
water Monitor ing Review, Vol. 2, pp. 35-40.
Green, R. E., L. R. Ahuja, and S. K. Chong, 1986.
"Hydraulic Conductivity, Diffusivity, and
Sorbivity of Unsaturated Soils: Field Methods." In:
Methods of Soils Analysis, Part I, Physical and
Mineralogical Methods, A. Klute, Editor. Agron-
omy, No. 9, Part 1, American Society of Agronomy,
Inc., Soil Science Society of America, Inc., Madison,
WI, pp. 799-823.
Gringarten, A.C., 1984. "Interpretation of Tests in
Fissured and Multilayered Reservoirs with Double
Porosity Behavior: Theory and Practice." Journal
of Petroleum Technology, April, 1984, pp. 549-564.
Grisak, G., W. Merritt, and D. Williams, 1977. "A
Fluoride Borehole Dilution Apparatus for Ground-
water Velocity Measurements." Canadian Geo-
tecknicalJournal, Vol. 14, No. 4, pp. 554-561.
Halvely, E., et al., 1966. "Borehole Dilution
Techniques - A Critical Review." In: International
Atomic Energy Agency, Isotopes in Hydrology,
Vienna.
Hantush, M. S., 1960. "Modification of the Theory of
Leaky Aquifers." Journal of Geophysical Research,
Vol. 65, No. 11, pp. 3713-3725.
Hantush, M. S., 1964. "Hydraulics of Wells." In:
Advances in Hydroscience, Vol. 1, Academic Press,
Inc., New York, NY, pp. 281-432.
Hvorslev, M. J., 1951. "Time Lag and Soil
Permeability in Ground-water Observations." U.S.
Army Corps of Engineers, Waterways Experiment
Station Bulletin 36, Vicksburg, MS.
Jorgensen, D. G., 1980. "Relationships between
Basic Soils-engineering Equations and Basic
Ground-water Flow Equations." U.S. Geological
Survey Water-Supply Paper 2064, 40 p.
Keely, J. F., 1982. "Chemical Time-series Sampling."
Ground-water Monitoring Review, Fall 1982 pp
29-38.
Keely, J .F. and K. Boateng, 1987. "Monitoring Well
Installation, Purging, and Sampling Techniques.
Part II: Case Histories." Groundwater, Vol. 25, pp.
427-439.
Keely, J. F. and F. Wolf, 1983. "Field Applications of
Chemical Time-series Sampling." Ground-water
Monitoring Review, Fall, 1983, pp. 26-33.
Kelly, W. E., 1976. "Geoelectric Sounding for
Delineating Ground-water Contamination."
Groundwater, Vol. 14, pp. 6-10.
Kerfoot, W. B., 1982. "Comparison of 2-D and 3-D
Ground-water Flowmeter Probes in Fully
Penetrating Monitoring Wells." In: Proceedings of
the Second National Symposium on Aquifer
Restoration and Ground-water Monitoring, D. M.
Nielson, Editor. Worthington, OH, Water Well
Journal Publishing Co., pp. 264-269.
Kerfoot, W. B. and S. M. Skinner, 1980. "Direct
Ground-water Flow Measurement." In: Proceed-
ings of 5th National Ground Water Quality
Symposium, Las Vegas, October 1980.
Keys, S. W. and L. M. MacCary, 1971. "Application
of Borehole Geophysics to Water Resources
Investigations." Techniques of Water Resources
Investigations of the United States Geological
Survey, Chapter El, Book 2.
Klute, A. and C. Dirksen, 1986. "Hydraulic Con-
ductivity and Diffusivity: Laboratory Methods." In:
Methods of Soil Analysis, Part I, Physical and
Mineralogical Methods, A. Klute, Editor. Agron-
omy, No. 9, Part 1, American Society of Agronomy,
Inc., Soil Science Society of America, Inc., Madison,
WI, pp. 687-734.
Kruseman, G. P. and N. A. de Ridder, 1970.
"Analysis and Evaluation of Pumping Test Data."
Bulletin 11, International Institute for Land
Reclamation and Improvement, Wageningen, The
Netherlands.
Lee, D. R., 1977. " A Device for Measuring Seepage
Flux in Lakes and Estuaries." Limnology and
Oceanography, Vol. 22, No. 1, pp. 140-147.
Lee, D. R. and J. A. Cherry, 1978. "A Field Exercise
on Groundwater Flow Using Seepage Meters and
Mini-piezometers." Journal of Geological Educa-
tion, Vol. 27, pp. 6-10.
Lohman, S. W., 1972. "Ground-water Hydraulics."
U.S. Geological Survey, Professional Paper 708.
Ludwig, K. J., 1983. "Electromagnetic Induction
Methods for Monitoring Acid Mine Drainage."
Ground-water Monitoring Review, Vol. 4 pp 46-
57.
MacKay, D. M., et al., 1986. "A Natural Gradient
Experiment on Solute Transport in a Sand
Aquifer. 1. Approach and Overview of Plume
Movement." Water Resources Research, Vol 22
No. 13, pp. 2017-2030.
Masch, F. D. and K. J. Denny, 1966. "Grain-size
Distribution and Its Effect on the Permeability of
Unconsolidated Sands." Water Resources Research
Vol. 2, pp. 665-677.
McLinn, E. L. and C. D. Palmer, 1988. "An
Electrical-resistivity Borehole-dilution Device for
the Determination of Ground-water Flux." Pro-
ceedings of the Second National Outdoor Confer-
ence on Aquifer Restoration, Ground-water Moni-
toring, and Geophysical Methods, Las Vegas NV
May 23-26,1988, pp. 851-874.
Mercer, J. W., S. D. Thomas, and B. Ross, 1982.
"Parameters and Variables Appearing in
38
-------
Repository Siting Models." U.S. Nuclear Regula-
tory Commission, Report NUREG/CR-3066.
Molz, F., et al., 1986. "Performance, Analysis and
Simulation of a Two-well Tracer Test at the Mobil
Site," Vol. 22, pp. 1031-1037.
Molz, F. J., et al., 1986. "Performance and Analysis
of Aquifer Tracer Tests with Implications for
Contaminant Transport Modeling." U.S. Envi-
ronmental Protection Agency, Research and
Development Report 600/2-86/062.
Morrison, R. D., 1983. Ground-water Monitoring.
Technology, Procedures, Equipment, and Applica-
tions. TIMCO Mfg., Inc., Prairie du Sac, WI.
Mualem, Y., 1986. "Hydraulic Conductivity of Un-
saturated Soils: Prediction and Formulas." In:
Methods of Soil Analysis, Part I, Physical and
Mineralogical Methods, A. Klute, Editor. Agron-
omy, No. 9, Part 1, American Society of Agronomy,
Inc., Soil Science Society of America, Inc.,
Madison, WI, pp. 799-823
Neuman, S. P., 1972. "Theory of Flow in Unconfined
Aquifers Considering Delayed Response of the
Water Table." Water Resources Research, Vol. 8,
pp. 1031-1045.
Neuman, S. P. and P. A. Witherspoon, 1972. "Field
Determination of the Hydraulic Properties of
Leaky Multiple-aquifer Systems." Water Resources
Research, Vol. 8, pp. 1284-1298.
Neuman, S. P., 1975. "Analysis of Pumping Test
Data from Anisotropic Unconfined Aquifers
Considering Delayed Gravity Response." Water
Resources Research, Vol. 11, No. 2, pp. 329-342.
Neuman, S. P., 1979. "Perspective on Delayed
Yield." Water Resources Research, Vol. 15, No. 4,
pp; 899-908.
Nguyen, V. and G. F. Finder, 1984. "Direct Calcula-
tion of Aquifer Parameters in Slug Test Analysis."
In: Ground Water Hydraulics, J. S. Rosenshein and
G. D. Bennett, Editors. Water Resources Mono-
graph Series 9, American Geophysical Union,
Washington, DC, pp. 222-239.
Palmer, C. D. and R. L. Nadon, 1986, "A Radial
Injection Tracer Experiment in a Confined
Aquifer, Scarborough, Ontario." Groundwater,
Vol. 24, pp. 322-331.
Palmer, C. D. and D. G. Paul, 1987. "Problems in the
Interpretation of Slug Test Data from Fine-
grained Glacial Tills." Proceedings of the Focus
Conference on Northwestern Ground Water Issues,
Portland, OR, May 5-7,1987.
Paul, D. G., 1987. "The Effect of Construction,
Installation, and Development Techniques on the
Performance of Monitoring Wells in Fine-grained
Glacial Tills." M.S. Thesis, Department of Geo-
logical and Geophysical Sciences, University of
Wisconsin-Milwaukee.
Pickens, J. F. and G. E. Grisak, 1981. "Scale
Dependent Dispersion in a Stratified Aquifer."
Water Resources Research, Vol. 17, pp. 1191-1212.
Rehm, B. W., et al., 1985. "Field Measurement
Methods for Hydrogeologic Investigations: A
Critical Review of the Literature." Electrical
Power Research Institute Report EA-4301, Project
2485-7, Electrical Power Research Institute, Palo
Alto, CA.
Rogers, R. B. and W. F. Kean, 1980. "Monitoring
Groundwater Contamination at a Fly-ash Disposal
Site Using Surface Electrical Resistivity
Methods." Groundwater, Vol. 18, pp. 472-478.
Scalf, M. R., et al., 1981. "Manual of Ground-water
Sampling Procedures." Robert S. Kerr Environ-
mental Research Laboratory, U.S. Environmental
Protection Agency, Ada, OK, pp. 43-71.
Slaine, D. D. and J. P. Greenhouse, 1982. "Case
Studies of Geophysical Contaminant Mapping at
Several Waste Disposal Sites." In: Proceedings of
the Second National Symposium on Aquifer
Restoration and Ground Water Monitoring, D. M.
Nielsen, Editor. National Water Well Association,
Columbus, OH, pp. 299-315.
Stewart, M. T., 1982. "Evaluation of Electromagnetic
Methods for Rapid Mapping of Salt-water
Interfaces in Coastal Aquifers." Groundwater, Vol.
20, pp. 538- 545.
Stollar, R. L. and P. Roux, 1975. "Earth Resistivity
Surveys - A Method for Defining Ground-water
Contamination." Groundwater, Vol. 13, pp. 145-
150.
Stolzenburg, T. R. and D. G. Nichols, 1985. "Prelim-
inary Results on Chemical Changes in Ground-
water Samples Due to Sampling Devices." Electric
Power Research Institute Report, EPRI EA-4118.
Sudicky, E. A., J. A. Cherry and E. O. Frind. 1983.
"Migration of Contaminants in Ground Water at a
Landfill: A Case Study. 4. A Natural-gradient
Dispersion Test." Journal of Hydrology, Vol. 63,
No. 1/2, pp. 81-108.
Sudicky, E. A., 1986. "A Natural Gradient
Experiment in a Sand Aquifer: Spatial Variability
of Hydraulic Conductivity and Its Role in the
Dispersion Process." Water Resources Research,
Vol. 22, No. 13, pp. 2069-2082.
Sutton, P. A. and J. F. Barker, 1985. "Migration and
Attenuation of Selected Organics in a Sandy
Aquifer — A Natural Gradient Experiment."
Groundwater, Vol. 23, No. 1, pp. 10-16.
Taylor, K., F. Molz, and J. S. Hayworth, 1988. "A
Single Well Electrical Tracer Test for the
39
-------
Determination of Hydraulic Conductivity and
Porosity as a Function of Depth." Proceedings of
the Second National Outdoor Action Conference on
Aquifer Restoration, Ground-water Monitoring
and Geophysical Methods, pp. 925-938.
Theis, C. V. 1935. "The Relation Between the Lower-
ing of the Piezometric Surface and the Rate and
Duration of Discharge of a Well Using Ground-
water Storage." Transactions of the American
Geophysical Union, Vol. 2, pp. 519- 524.
Walton, W. C., 1962. "Selected Analytical Methods
for Well and Aquifer Evaluation." Illinois State
Water Survey Bulletin 49.
Walton, W. C., 1970. Groundwater Resource Evalua-
tion. McGraw-Hill, New York, NY.
Walton, W. C., 1984. Practical Aspects of Ground
Water Modeling, National Water Well Associ-
ation, Dublin, OH.
Way, S. C. and C. R. McKee, 1982. "In-situ Deter-
mination of Three-dimensional Aquifer Perme-
abilities." Groundwater, Vol. 20, pp. 594-603.
Weeks, E. P., 1969. "Determining the Ratio of
Horizontal to Vertical Permeability by Aquifer-
test Analysis." Water Resources Research, Vol. 5
No. 1, pp. 196-214.
40
-------
CHAPTER 5
SUBSURFACE CHEMICAL PROCESSES
Richard L. Johnson, Carl D. Palmer, and William Fish
Introduction
Risk assessment and remediation of ground-water
contamination require an understanding of how
chemicals move through and interact with the
subsurface environment. However, subsurface
transport of contaminants is often controlled by
complex interactions between chemical, physical and
biological processes. This means that mathematical
models often must be used to predict chemical
movement. Models for ground-water systems are
frequently based on the advection-dispersion
equation described in Chapter 2 and shown here in
one-dimensional form:
D
aC
ax
- v —
2 ax
at
(D
This equation requires quantitative information to
simulate all subsurface processes. In the context of
the advection-dispersion equation, chemical
processes important in controlling contaminant
transport and fate usually take the form of reaction
terms (RXN) added to the basic equation:
D
a2c
ac
_ v _ =
ax
ac
_
at
(2)
This chapter will discuss the chemical processes that
affect the movement of organic and inorganic
contaminants in ground-water systems and how
those processes can be represented in the advection-
dispersion equation. Methods and limitations of
experimental data for modeling chemical processes
also will be examined.
Reactions of Organic Compounds
Hundreds of thousands of organic chemicals are
currently used in industrial and domestic
applications in the United States. These chemicals
represent an extremely broad range of physical and
chemical properties and are subject to different
physical, chemical, and biological processes in the
subsurface. For example, because chlorinated
solvents are only slightly soluble and are more dense
than water, they can penetrate deep into aquifers
and remain as an immiscible phase for prolonged
periods of time. In contrast, a spill of acetone will not
penetrate into the ground water because of its low
density. Also, because of its miscibility with water,
the acetone will dissolve quickly and become
available for further chemical and biological
reactions.
Organic reactions may transform one compound into
another, change the state of a compound, or cause a
compound to combine with other organic or inorganic
chemicals. In the context of the advection-dispersion
equation, these reactions represent changes in the
distribution of mass within the elementary volume
through which the movement of the chemicals is
modeled. Although, many of these reactions have
been studied in the laboratory and observed in the
field, there is a lack of good quantitative information
about the processes under complex, real-world
conditions.
Chemical reactions in the subsurface are frequently
characterized on a kinetic basis as equilibrium or
zero- or first-order, depending upon how the rate is
affected by the concentrations of the reactants. For
example, a zero-order reaction is one that proceeds at
a rate independent of the concentration of the
41
-------
reactant(s). In a first-order process, the rate of the
reaction is directly dependent on the concentration of
one of the reactants. In many cases, grouping a
chemical process into one of these classes over-
simplifies the actual process. However, more
realistic, higher-order processes often are difficult to
measure and/or model in complex environmental
systems. The following example of hydrolysis
illustrates some of the difficulties in obtaining
quantitative kinetic information and applying it to
the advection- dispersion equation.
Hydrolysis
Hydrolysis is the direct reaction of dissolved
compounds with water molecules, and can be an
important abiotic degradation process in ground-
water systems (Mabey and Mill, 1978). For example,
hydrolysis of chlorinated hydrocarbons is significant
because many chlorinated compounds are not readily
degraded by reactions such as biodegradation
(Siegrist and McCarty, 1987). This hydrolysis of
chlorinated compounds often yields an alcohol or an
alkene (Figure 34).
RX + HOH
HX
ROH + HX
I I
C-C
Figure 34.
OC + HX
Schematic hydrolysis reactions for halo-
carbons.
Hydrolysis rates are typically difficult to measure in
the field due to other competing reactions and slow
degradation rates, and as a result, most hydrolysis
data come from laboratory studies. Even in the
laboratory, conditions must be optimized to minimize
other reactions and provide good quantitative data.
For example, data in Figure 35 (Ellington, et al.,
1986) for the hydrolysis of 1,2,4-trichlorobenzene
were collected at 70°C to speed the reaction and
eliminate competing microbial degradation. The
data then were fitted as a first-order reaction, which
assumes a straight line on a semi-log plot, even
though the data suggest some higher-order effects.
From these data, the hydrolysis rate constant (K) at
the elevated temperatures was estimated (Figure
36). The extrapolation of K at ambient temperatures
100 Q
50
PERCENT
REMAINING
10
70°C
pH"7.11
HALF-LIFE-160 HOURS
50 100 150
TIME (HOURS)
.Figure 35. Hydrolysis data for 1,2,4-trichlorobenzene
(adapted from Ellington, et al., 1986).
dC
In
CCO)
-Kt
at the half-life:
C(0)
0.5
t s 160 hours
thus, K = 0.69/160
K =4.3x 103 hr"1
Figure 36. Calculation of the first-order hydrolysis rate
constant for the degradation of 1,2,4-
trichlorobenzene.
from this data could potentially result in additional
error.
After estimating K, the behavior of a specific
compound can be modeled using a form of the
advection-dispersion equation that includes a first-
order degradation term:
„
D
a2c
_v
sc
= _ +KC
(3)
42
-------
Although first-order reactions are not always the
most realistic, they are easy to incorporate into
transport models. Hydrolysis and other chemical
process data on the EPA priority pollutants can be
found in Mabey, et al. (1982).
Sorption
Of all the chemical processes discussed in this
chapter, sorption is probably the most important and
the most studied. In ground-water applications,
sorption of non-polar organics often can be treated as
an equilibrium-partitioning process between the
aqueous phase and the porous medium (Chiou, et al.,
1979). When solute concentrations are low (e.g.,
either ^ 10-5 Molar, or less than half the solubility,
whichever is lower), partitioning often is described
using a linear Freundlich isotherm (Karickhoff, et
al. 1979; Karickhoff, 1984):
where
S = K C
P
(4)
where S is the sorbed concentration (mg/g), C is the
aqueous concentration (mg/mL), and Kp is the
partition coefficient. Kp typically is measured in the
laboratory using batch equilibrium tests, and the
data are plotted as the concentration in the aqueous
phase versus the amount sorbed onto the solid phase
(Figure 37) (Chiou, et al., 1979). Under conditions of
o
V)
1200
o>
O
<
oc
I
o
o
S
m
oc
o
W
800
400
1,1,1 -TRICHLOROETH ANE
1,1,2,2-TETR ACHLOROETH ANE
1,2-DICHLOROETHANE
0 400 800 1200 1600 2000 2400
AQUEOUS CONCENTRATION (ug/U
Figure 37. Batch equilibrium data for 1,1,1-TCA, 1,1,2,2-
TeCA and 1,2-DCA (adapted from Chiou, et
al., 1979).
linear equilibrium partitioning, the sorption process
can be represented in the advection-dispersion equa-
tion as a "retardation factor," R:
32c ac ^ac
D— -v — =R —
ax2 *< at
(5a)
,K
R= 1 +
(5b)
and pb = soil bulk density (g/cm3) and 0t = porosity.
The dominant mechanism of organic sorption is the
hydrophobic bond between a chemical and natural
organic matter associated with aquifers (Karickhoff,
et al., 1979; 1984; Tanford, 1973; MacKay and
Powers, 1987; Chiou, et al., 1985). The extent of
sorption can be reasonably estimated if the organic
carbon content of the soil is known (Figure 38)
(Karickhoff, et al., 1979; Karickhoff, 1984) by using
the expression:
K =K f
p oc oc
(6)
where foc is the fraction organic carbon content of the
soil and Koc is a proportionality constant character-
istic of the specific chemical. This approach works
reasonably well for a wide range of soils, providing
the soil organic content is sufficiently high (e.g., foc
> 0.001). For lower carbon-content soils, sorption of
the neutral organics onto the mineral phase can
cause important errors in the estimate of Kp (Chiou,
et al, 1985). Still, the Koc relationship allows estima-
tions of sorption-based retardation developed from
measured foc values, rather than the more expensive;
batch equilibrium tests.
1800
1500
1200
PYRENE
900
600
300
SLOPE • K
OO
'PHENANTHRENE
600
500
400
Kp
300 PHENAN-
THRENE
200
100
0.0 .005 .010 .015 .020 .025
FRACTION ORGANIC CARBON
Figure 38. Sorption of pyrene and phenanthrene on
various soils as a function of soil foc (after
Karickhoff, 1981).
Koc values for many compounds are unknown.
Because of this, numerous researchers have
developed correlation equations to relate Koc to more
commonly available chemical properties, such as
solubility or octanol-water partition coefficient
(Chiou, etal., 1982 and 1983; Schwarzenbach and
Westall, 1981; Kariekhoff, 1981; Kenaga and Goring,
43
-------
1980) (Figures 39 and 40). Some of these correlations
cover a broad range of compounds, in which case the
errors associated with the Koc estimates can be large.
However, within a compound class, Koc values
derived using these expressions often provide
reasonable estimates of sorption.
7
6
5
4
3
2
1
0
-1
log Koc * -0.55 log S + 3.64
(S in mg/L )
Kenaga and Goring, 1980
D a
-4 -3-2-101 2
log S (mg/L)
4 5
Figure 39. Correlation of Koc and solubility data
(adapted from Kenaga and Goring, 1980).
Regression Equations
tog KOO - -0.55 log S + 3.64
log Koc = 0.544 log Kow + 1.377
tog Koc = 0.681 log BCF + 1.963
Figure 40. Regression equations for Koc versus solu-
bility (S in mg/L), octanol/water partition
coefficient (Kow), and bioconcentration
factor (BCF) (after Kenaga and Goring, 1980).
Unfortunately, the linear equilibrium approach to
sorption is not adequate for some real-world
situations. For example, trichloroethylene (TCE)
sorption onto a glacial till shows a change in Kp of
more than 50-fold over the parts per billion range
(Figure 41) (Johnson, et al., 1989; Myrand, et al.,
1989; McKay and Trudell, 1987). In addition, for
many very hydrophobia organics, adsorption and
desorption occur over time scales of many months
(Figure 42) (Karickhoff and Morris, 1985;
Witkowski, et al., 1988; Coates and Elzerman, 1986;
Wu and Gschwend, 1986). The importance of non-
ideal sorption will be discussed again in Chapter 6.
Field data or field-scale experiments also can provide
good measures of sorption-based retardation. Field
data have the advantages of operating on more
realistic scales, time frames, and conditions than
those typically reproduced in the laboratory. For
r McKay and Trudell
LOG OF THE FINAL AQUEOUS CONCENTRATION (PPB)
Figure 41. TCE sorption on a glacial till from near
Sarnia, Ontario (after McKay and Trudell,
1987; Myrand, et al. 1989; Johnson, et al.,
1989).
DESORPTION OF HEXACHLOROBENZENE
Figure 42. Desorption of hexachlorobenzene from two
sediments (after Karickhoff and Morris,
1985).
example, data on chlorophenol transport collected
downgradient from a hazardous chemical disposal
site at Alkali Lake, Oregon (Figure 43) (Johnson, et
al., 1985), demonstrate compound-specific retarda-
tion that can be explained by equilibrium-parti-
tioning sorption. Also, numerous large-scale tracer
experiments were conducted in recent years to
demonstrate and quantify certain physical and
chemical processes (Figure 44) (LeBlanc, et al., 1987;
MacKay, et al., 1986; Barker, et al., 1987; Molz, et
al., 1986). However, field data are often expensive, as
well as difficult and time-consuming to gather.
Cosolvation and lonization
Cosolvation and ionization are processes that can
decrease sorption and, therefore, increase transport
velocity. Cosolvents decrease the entropic forces that
favor sorption of hydrophobic organics by increasing
the interactions between the solute and the solvent
(Nkedi-Kizza, et al., 1985; Zachara, et al., 1988). The
thermodynamic basis for the Cosolvation effect was
described by Rao, et al. (1985), and Woodburn, et al.
(1986). For many of the more hydrophobic priority
44
-------
pollutants, the presence of biologically derived or
anthropogenic compounds in the range of 20 percent
by volume or greater can increase the solubility of
those pollutants by an order of magnitude or more.
As seen in Figure 45, this results in a log-linear
decrease in sorption that inversely parallels changes
in the solubility of the solute in the mixed solvent
system (Nkedi-Kizza, et al., 1985). The data for three
soils in Figure 45 clearly show that the cosolvent
concentration must be large in order for the solute
velocity to be substantially increased. For this
reason, cosolvation is important primarily near
sources of ground-water contamination.
ALKALI LAKE
2,3,4,5-TETRACHLOROPHENOL
DICHLOROPHENOXYPHENOL
200
400
Figure 43.
DISTANCE (M)
Chlorophenol distributions dow.ngradient of
the chemical disposal site at Alkali Lake,
Oregon (adapted from Johnson, et al., 1985).
STANFORD/WATERLOO
CARBON TETRACHLORIDE
TETRACHLOROETHYLENE
200
400
Figure 44.
TIME (DAYS)
Tracer distribution data from the Stanford/
Waterloo tracer experiment (adapted from
Mackay, et al., 1986).
Acidic compounds, such as phenols or organic acids,
can lose a proton in solution to form anions (Figure
46) that, due to their charge, tend to be very water
soluble (Zachara, et al., 1986). Thus, the Koc of a
compound like 2,4,5-trichlorophenol can decrease
from 2,330 for the phenol, to near zero for the
phenolate (Figure 47). Acidic compounds tend to
ionize more as the pH increases. However, for many
compounds like the chlorophenols, substantial ion-
ization can occur at neutral pHs. Positively charged
organic ions also can be present in ground-water
systems. Zachara, et al. (1986) and Ainsworth, et al.
(1987) demonstrated that in low-carbon soils and
1000
100
Kp 10
0.1
ANTHRACENE
.2 .3 .4 .5
Figure 45.
FRACTION CO-SOLVENT
(METHANOL)
Effect of methanol as a cosolvent on anthra-
cene sorption for three soils (Adapted from
Nkedi-Kizza, et al., 1985).
Koc-2330
Koc~°
Figure 46. Koc values for 2,4,5-trichlorophenol and 2,4,5-
trichlorophenolate.
45
-------
2500
2000
KOC 1500
1000
500
2,4,5-
TRICHLOROPHENOL
(Roberts, et al., 1986) show a reasonably good fit to a
first-order degradation process.
6.0 6.5 7.0 7.5 8.0 8.5
P"
Figure 47. Koc versus pH for 2,4,5-trichlorophenol.
soils containing clays, quinoline is sorbed primarily
by ion exchange.
Zachara, et al. (1988) and Fu and Luthy (1986a,
1986b) examined the combined processes of ioniza-
tion and cosolvation using quinoline as a model.
They found that sorption decreased in a log-linear
fashion when methanol and acetone were used as
cosolvents. Zachara, et al. were able to relate the
increased cosolvent effect of acetone, relative to
methanol, to the greater solvating properties of
acetone.
Biodegradation
Although biodegradation is described in greater
detail in Chapters 7 and 8, two chemical-reaction-
related aspects of biodegradation are discussed in
this section. The first aspect is that biodegradation in
natural systems often can be modeled as a first-order
chemical reaction. Both laboratory and field data
suggest that this is true when none of the reactants
are in limited supply. For example, data from the
Stanford/Waterloo tracer experiment (Figure 48)
Cl
Cl
oc
Cl
Cl
Cl
Cl
c-c
TETRACHLOROETHYLENE
• X
CARBON TETRACHLORIDE
A
BROMOFORM
DIOHLOROBENZENE
200 400 600
TIME (DAYS)
Figure 48. Degradation of chlorinated organics in the
Stanford-Waterloo natural-gradient tracer
experiment (adapted from Roberts, et al.,
1986).
The second chemically related aspect of biodegra-
dation is that, like most abiotic reactions, not all
biodegradation reactions result in complete mineral-
ization of the reactants. In many cases, the micro-
organisms that degrade the contaminants produce
an intermediate chemical which they do not or
cannot degrade. Other organisms or other conditions
may allow degradation to continue, but the lifetime
of the intermediate chemical may be significant.
Perhaps the best known example of this is the
sequential degradation of tetrachloroethylene to the
more toxic vinyl chloride (Figure 49) which is not
readily degraded and tends to accumulate. In the
process, the original moderately sorbing contam-
inant is transformed to a weakly sorbing compound
with a partition coefficient lower than the original
chemical by a factor of approximately eighty.
Volatilization and Dissolution
Spills of volatile organic compounds (solvents or
petroleum-based liquids) are very common in all
parts of the United States. Two important transport
364 126 59
Figure 49. Schematic of the sequential degradation of tetrachloroethylene to vinyl chloride.
8.2
46
-------
pathways for these materials are volatilization into
the unsaturated zone and dissolution into the ground
water. In addition to being important transport
pathways, volatile organic compounds also carry the
contaminants from the free-product phase into the
aqueous and vapor phases where they are more
amenable to degradation.
The importance of volatilization is determined by the
area of contact between the free product and the
unsaturated zone, the vapor pressures of the spilled
compounds, and the rate at which the compound
diffuses in the subsurface. The contact between a
compound and the unsaturated zone is determined
by the nature of the medium (e.g., grain size, depth to
water, water content, etc.) as well as of the compound
(e.g., surface tension and liquid density). As
discussed in Chapter 3, when immiscible liquids
move downward through unsaturated porous media,
portions of the liquid are left behind as "trapped
residual." This residual provides a very large surface
area for volatilization. Laboratory experiments
suggest that vapor concentrations in the vicinity of
the residual are maintained at saturation
concentrations. Movement of the vapor away from
the residual is typically controlled by molecular
diffusion, as described by Fick's Second Law:
eax2
at
(7)
Mass transfer is controlled by the effective diffusion
coefficient, De, of the compound in the porous
medium. De is defined as:
D =D T
e a
(8)
where Da = free-air diffusion coefficient, and T; =
tortuosity factor. Tortuosity factors for moist porous
media can be determined experimentally or
calculated theoretically. Millington (1959) developed
the most widely used theoretical expression for T;:
0
2.33
T =
e
(9)
where 9a is the air-filled porosity of the medium.
Thus, diffusion is significantly reduced in high
water-content (low 9a) soils (Figure 50).
In addition to the effect of tortuosity, diffusion may
be further reduced by the partitioning of the vapors
out of the gas phase and into the solid or aqueous
phases. This process can be described in a manner
that is very analogous to retardation due to sorption
in the saturated zone, with the addition of a term to
describe the partitioning between the vapor and
aqueous phases (Baehr, 1987):
PbKHKP
Figure 50.
0.3
Tortuosity as a function of air-filled porosity
(total porosity = 0.35) (adapted from
Millington, 1959).
where KH = dimensionless water-air partition
coefficient and 9W = volumetric pore-water content.
Equation 7 can be modified to the form:
R
ac
at
(ii)
As with the saturated zone R (Equation 5b), this
formulation of the retardation factor assumes that
partitioning is at equilibrium and the factors Kn and
KH are not functions of solute concentration. These
assumptions are probably valid for most unsaturated
zone conditions.
When immiscible fluids reach the capillary fringe,
their behavior is dictated by the fluids' density
relative to water (Schwille, 1988; Scheigg, 1984). As
discussed in Chapter 3, fluids less dense than water
(LNAPLs) pool up on the water table while dense
fluids (DNAPLs) penetrate into the ground water.
Floating pools of LNAPL also can provide sub-
stantial surface area for volatilization. Again,
diffusion frequently controls the mass transfer of
organics into the vapor phase.
The transport and fate of DNAPLs that penetrate
into the ground-water zone is controlled by
dissolution. Anderson (1988) and Anderson, et al.
(1987) conducted a series of dissolution experiments
47
-------
in a three-dimensional physical model (Figure 51). A
cylindrical volume of sand laden with tetrachloro-
ethylene (TeCE) was surrounded by identical sand
without the TeCE. Ground water flowing through
the tank dissolved the TeCE, and concentrations
downgradient of the source quickly rose to saturation
values. Experiments at higher ground-water
velocities showed that saturation values were
CYLINDRICAL SPILL OF TeCE
QROUNDWATER,
FLOW £f
EXPERIMENTAL SAND TANK
Figure 51.
1m x 1m x 1m
Schematic of the experimental apparatus for
measuring dissolution from DNAPL residual in
saturated porous media used by Anderson
(1988).
maintained, even when the flow rate was 1 m/day
(Figure 52). These data suggest that ground water is
flowing relatively unimpaired through the zone of
residual, and that the dissolution process should be
effective, even at the high ground-water velocities
present during remediation.
Chemical Reactions of Inorganic
Components
Spec/at/on
For organic contaminant studies, researchers are
primarily interested in the total concentration of the
compound in a given phase (e.g., in water vs. in the
aquifer matrix). Study of inorganic compound
behavior, on the other hand, is greatly complicated
by the lack of sufficient knowledge of the total
concentration of material. Inorganic materials can
occur in many chemical forms, and knowing these
forms or "species" is critical to predicting the
behavior of inorganic compounds (Morel, 1983;
Sposito, 1986).
In ground water, an element may occur in any of the
following six categories of species:
10 cm/day
30 cm/day
100 cm/day
o
,§
U!
o
0)
200
180
160
140
120
100
80
60
40
20
Figure 52.
-20-16-12-8 -40 4 8 12 16 20
DISTANCE FROM PLUME CENTER (cm)
Tetrachloroethylene dissolution data for ground-water velocities of 10 cm/day, 30 cm/day, and 100
cm/day (adapted from Anderson, et al., 1987).
48
-------
1. "Free" ions (i.e., surrounded only by water
molecules)
2. Insoluble species (e.g.,Ag2S,BaSO4)
3. Metal/ligand complexes (e.g., A1(OH)2 + , Cu-
humate)
4. Adsorbed species (e.g., lead sorbed onto a ferric
hydroxide surface)
5. Species held on a surface by ion exchange (e.g.,
calcium ions on clay)
6. Species that differ by oxidation state (e.g.,
manganese (II) and (IV); iron (II) and (III); and
chromium (III) and (VI)
The mobility, reactivity, biological availability, and
toxicity of metals and other inorganics depend upon
the speciation; knowing only the total concentration
of an inorganic compound is frequently of little use.
The primary reactions governing these six categories
of inorganic chemicals are discussed below.
Solubility and Dissolution
The dissolution and weathering of minerals deter-
mines the natural composition of ground water. A
useful distinction can be drawn between these two
related phenomena. "Dissolution" refers to the
dissolving of all components within a mineral, for
example:
halite (NaCl) -> Na + , Cl-
gypsum (CaSO4 • 2H2O) ->Ca2+, SO4=
calcite/aragonite (CaCOailimestone) -»Ca2 + , COa",
HCO3-
quartz (SiO2) -» H4SiO4
Dissolution of such minerals is the source of most
inorganic ions in ground water. The extent of the
dissolution can be estimated from calculations using
thermo'dynamic constants known as the solubility
products, Ksp. In principal, a mineral can dissolve up
to the limits of its solubility; however, in many cases,
the reactions occur at such a slow rate that true
equilibrium is never attained (Morgan, 1967).
Natural systems are further complicated by the
simultaneous presence of many minerals containing
common ions. The contribution of ions from one
mineral affects the solubility of other minerals
containing the same ion. This is the so-called
"common-ion effect." Computer calculation schemes,
such as MINTEQ (Felmy, et al., 1984), MINEQL
(Westall, et al., 1976), or WATEQ2 (Ball, et al.,
1980), yield the equilibrium distribution of chemical
species in the ground water and indicate if the water
is undersaturated, supersaturated, or at equilibrium
with various mineral phases. Some of these pro-
grams also can be used to predict the ionic
composition of ground water in equilibrium with
assumed mineral phases (Jennings, et al., 1982).
"Weathering" is a partial dissolution process in
which certain elements leach out of a mineral,
leaving others behind. The weathering of alumino-
silicates (such as feldspars) contributes cations,
primarily Ca2 + , Mg2 + , K + , Na + , and silica, to
water and forms secondary weathering products such
as kaolinite and montmorillonite. This weathering of
silicate minerals increases the alkalinity of the
water; hence, natural ground water often is more
alkaline than its rainwater origins. Weathering and
dissolution also can be sources of contaminants.
Leachates from mine tailings (Hem, 1970) can yield
arsenate, heavy metals, and strong mineral acids.
Also, leachates from fly-ash piles yield selenium,
arsenate, lithium ions (Li + ), and heavy metals
(Honeyman, et al., 1982; Murarka and Macintosh,
1987; Stumm and Morgan, 1981).
The converse of dissolution reactions is the pre-
cipitation of minerals or contaminants out of
aqueous solution. During precipitation, the least-
soluble mineral is removed from solution as shown
for iron in Figure 53 (Stumm and Morgan, 1981;
Williams, 1985). A thermodynamic restriction
known as the Gibbs Phase Rule limits the number of
solid phases that can form from a given solution
(Sillen, 1967). An element is removed by precipita-
tion when its solution concentration saturates the
solubility of one of its solid compounds. If the
solution concentration later drops below the
solubility limit, the solid will begin to dissolve until
the solubility level is attained again. Thus,
contaminants may initially precipitate, then slowly
dissolve later after "remediation" reduced the
Figure 53.
2 46 8 10 12 14
PH
Log C-pH diagram for iron (adapted from
Stumm and Morgan, 1981).
49
-------
solution concentration. Remediation may take years
to complete under such conditions.
In another scenario, a contaminant initially may be
soluble, but later precipitates after mixing with
other waters or after contact with other minerals
(Williams, 1985; Drever, 1982; Palmer, 1989). For
example, pumping water from an aquifer during
remediation might cause dissolved lead to be
mobilized until it converges and mixes with high
carbonate waters from a different formation. At that
point, much of the lead would precipitate as PbCOs
solid. Changes in the oxidation state of an element
also can cause contaminants to precipitate or
dissolve; this topic is addressed later in this chapter
in the discussion on redox chemistry.
Complexation Reactions
In a complexation reaction, a metal ion reacts with
an anion that functions as a so-called ligand. The
metal and the ligand bind together to form a new,
soluble species called a complex. Transition metals
are the most important metals involved in complexa-
tion (Stumm and Morgan, 1981); alkaline earth
metals only form weak complexes while alkali
metals essentially do not form complexes at all
(Dempsey, et al., 1983). The approximate order of
complexing strength of metals is:
Fe(III) > Hg > Cu > Pb > Ni > Zn > Cd > Fe(II)
> Mn > Ca > Mg
Important inorganic ligands include most of the
common anions (Hanzlik, 1976), and their strength
depends highly on the metal ion with which they are
complexing. Common ligands are OH-, C1-, SC>4 =,
CO3=, S=, F-, NH3, PO43-, CN-, and polyphosphates.
Inorganic ligands frequently are in great excess
compared to the "trace" metals they bind and,
therefore, affect metal chemistry, not vice versa
(Morel, 1983).
Organic ligands generally form much stronger
complexes than inorganic ligands. Important organic
ligands include synthetic compounds from wastes
such as amines, pyridines, phenols, and other
organic bases and weak acids. Natural organic
ligands are mostly humic materials (Stevenson, 1982
and 1985; Hayes and Swift, 1978; Schnitzer, 1969),
and the complexation behavior of these diverse
substances is difficult to predict (Dzombak, et al.,
1986; Fish, et al., 1986; Perdue, 1985; Sposito, 1984;
Perdue and Lytle, 1983). Humic materials are
generally found in significant concentrations only in
shallow aquifers, but in such systems they may
dominate the metal chemistry of the ground water
(Thurman, 1985).
Equilibrium among reactants and complexes for a
given reaction is predicted by an equilibrium (or
"stability") constant (K) which defines a mass-law
relationship among the species. For example:
Reaction: Hg2+ + Cl' = HgCl +
Described by: [HgCl + ]/[Hg2 + ][Cr] = KHgci+
= 107.2
For given total ion concentrations (measured by
analysis), stability constants can be used to predict
the concentration of all possible species (Figure 54)
(Smith and Martell, 1976).
pH
>- 7.19 7.2 7.5 8.0 8.29
18
\Ji
i 20
22
*
u>
o
0.1
1.0 10.0
SALINITY (0/00)
Figure 54. Speciation of mercury as a function of
salinity (adapted from Smith and Martell,
1976).
Because complexes decrease the amount of free ions
in solution, less metal may adsorb onto aquifer
matrix or precipitate. That is, the metal is more
soluble because it is mostly bound up in the soluble
complex. Rueter, et al. (1979) found that a metal
undergoing complexation may be less toxic to aquifer
microbes.
Adsorption and Surface Chemistry
A vast amount of surface area exists in an aquifer
and, in many cases, surface adsorption is the most
important process governing toxic metal transport in
the subsurface. Changes in metals concentration, as
well as pH, can have a significant effect on the extent
of adsorption (Figure 55). Unfortunately, a general
model of ionic adsorption on natural surfaces still
has not been developed (Dzombak, 1986; Dzombak
and Morel, 1986). Numerous theories and models of
adsorption exist, but no truly general principles have
been defined. Current models still are highly
50
-------
100
s
CO
80
60
< 40
20
Fe(lll)
Pb
Cd
4
PH
6
8
Figure 55.
Adsorption of metal ions on amorphous silica
as a function of pH (adapted from Schindler,
et al., 1976).
empirical and therefore are calibrated to specific
data (Westall, 1980). Because of this, it is difficult to
base generalized regulations or remediation plans on
the computer-predicted behavior of metal ions.
Despite these shortcomings, some useful approaches
to the problem have been developed. Adsorption data
usually are presented graphically as "isotherms"
(called this because they represent data collected at a
fixed temperature). Isotherms can be fitted to
mathematical representations; the two most common
forms are the Freundlich and the Langmuir
isotherms (Figure 56). The Freundlich isotherm:
iN
(12)
is purely empirical, and sorbed (S) and aqueous (C)
concentration data are fitted by adjusting the
parameters K and N. The Langmuir formulation, in
contrast, is based on the theory of surface
complexation (Morel, 1983):
S = S
KG
"maxl + KC
(13)
where Smax is the maximum amount which can be
sorbed, and K is the partition coefficient.
Surface complexation models represent adsorption as
ions binding to specific chemical functional groups
on a reactive surface. All surface sites may be identi-
cal or may be grouped into different classes of sites
(Benjamin and Leckie, 1981), and each type- of
surface site has a set of specific adsorption constants,
one for each adsorbing compound. Electrostatic
forces at the surface also contribute to the overall
adsorption constant (Davis, et al., 1978) and
sometimes are explicitly included in the adsorption
constant as the Coulombic term (Stumm and
Morgan, 1981). The binding of ions to the surface is
logS
logC
Figure 56.
Schematic drawing of Freundlich and
Langmuir isotherm shapes for batch
equilibrium tests.
computed from constants with mass-law equations
identical to those used to calculate solution-complex
formation (Stumm, et al., 1976; Schindler, et al.,
1976; Dzombak and Morel, 1986).
Although this approach is conceptually simple, there
are practical problems in using the models. Model
parameters are effectively data-fitting parameters.
The parameters can fit a specified set of data to a
particular model, but have no true thermodynamic
meaning and, therefore, no generality beyond the
calibrating data set (Westall, et al., 1980). Some
models contain many "knobs" that can be adjusted to
get the best fit of the experimental data. However,
once a model is calibrated for one set of conditions,
lack of generality may preclude dynamic predictions
for changing aquifer conditions. Parameters
calculated using one model should not be used in
another model unless modified according to each
model's assumptions.
Ion-Exchange Reactions
Ion-exchange reactions are similar in effect to
adsorption, but have some key distinctions.
Adsorption is viewed as the coordination bonding of
metals (or anions) to specific surface sites considered
to be two-dimensional. In contrast, an ion-exchanger
is visualized as a three-dimensional, porous matrix
containing fixed charges. Ions are held by
electrostatic forces rather than by coordination
51
-------
bonding (Helfferich, 1962). Surface complexation
uses fixed stability constants for mass law
calculations whereas ion-exchange "selectivity
coefficients" are strictly empirical (Reichenberg,
1966) and vary with the amount of ion present. Ion
exchange best describes the binding of alkali metals,
alkaline earths, and some anions to clays and
condensed humic matter (Sposito, 1984; Helfferich,
1962).
Knowledge of ion exchange in soils and aquifers is
very important for understanding the behavior of
major (natural) ions. In principle, ion exchange
theory also is useful for understanding low-level
contaminant ions. In practice, this theory is very
empirical when applied to these types of ions, and
surface complexation models probably are better
choices for trace metals. Ion exchange models need
further development, but may be the most useful
representation of competition among metals for
surface binding (Sposito, 1984).
Redox Chemistry
Reduction-oxidation (redox) reactions involve a
change in the oxidation state of elements. The level
of that change is determined by the number of
electrons on the element transferred during the
reaction (Stumm and Morgan, 1981.) Redox
reactions can greatly affect contaminant transport.
For example, in slightly acidic to alkaline
environments, Fe(III) precipitates as a highly
adsorptive solid phase (ferric hydroxide), whereas
Fe(II) is very soluble and does not retain other
metals. The reduction of Fe(III) to Fe(II), therefore,
releases not only Fe2+ to the water, but also any
contaminants that were adsorbed to the ferric
hydroxide surfaces (Evans, at al., 1983; Sholkovitz,
1985). The behavior of chromium and selenium also
illustrates the importance of redox chemistry to
contaminant movement. Cr(VI) (hexavalent) is a
toxic, relatively mobile anion whereas trivalent
Cr(III) is inert, relatively insoluble, and strongly
adsorbs to surfaces. Selenate (Se(VD) is mobile, but
less toxic; however, selenite (Se(IV)) is more toxic,
but less mobile.
The redox state of an aquifer usually is closely
related to the microbial activity and the type of
substrates available to the organisms. Organic
contaminants provide the reducing equivalents for
the microbes. Because of the inherently "enclosed"
nature of aquifers, oxygen is readily depleted and
chemically reducing (anaerobic) conditions form. The
redox reactions that occur depend entirely on the
dominant electron potential, which is defined by the
primary redox-active chemical species. For example,
the combination of Fe(II)/Fe(III) defines a particular,
narrow range of electron potentials, whereas
S( + IV)/S(-II) defines a more broad range. These
pairs of chemical species are called redox couples.
After oxygen is depleted from the water, the most
easily reduced materials begin to react and, along
with the reduced product, dictate the dominant
potential. After that material is more or less
completely reduced, the next most easily reduced
material begins to react, and so on. Microbes
typically catalyze this series of reactions, and an
aquifer is described as "mildly reducing" or "strongly
reducing," depending on where it is in the chemical
series (Stumm and Morgan, 1981).
The electron potential of a water body can be given in
volts (as the EH), or expressed by the "pe," which is
the negative logarithm of the electron "activity" in
the water. The pe sounds complex but is exactly
analogous to pH (negative log of hydrogen ion
activity). A convenient way to summarize a set of
reactions is on a pH-pe (or pH-En) diagram, some-
times called a Pourbaix diagram. Inspection of the
diagram yields the predominant redox species at any
specified pH and EH (Figure 57).
+ 1.0
+0.8
+0.6
+0.4
Eh +0-2
+0.0
-0.2
-0.4
-0.6
-0.8
-1.0
%* IX
*< xX,
Svftv. *->%
^
Figure 57.
02 468 10 12 141
pH
pH-Eh diagram showing the ranges of various
aquatic environments.
In this theoretical approach, only one redox couple
should define the redox potential of the system at
equilibrium. In practice, this series of events is not
52
-------
clearly defined and many redox couples not in
equilibrium can be observed simultaneously
(Lindberg and Runnels, 1984). Hence, it can be
nearly impossible to predict redox behavior of
chemicals in aquifers from equilibrium predictions.
However, the redox status of an aquifer cannot be
ignored because of the effects discussed above and
the potential effects on biodegradation of organic
contaminants. Anaerobic (reducing) conditions are
inefficient for the degradation of hydrocarbons, but
reducing conditions favor dehalogenation of chlorin-
ated and other halogenated compounds.
Much more research is needed in the area of redox
chemistry in aquifers because researchers only now
are realizing the importance of these phenomena in
controlling contaminant behavior. Among the areas
for future research are the development of better
measuring techniques for redox potentials, micro-
biological studies, and new approaches for computa-
tional models.
References
Ainsworth, C. C., J. M. Zachara, and R. L. Schmidt,
1987. "Quinoline Sorption on Na-Montmorillonite:
Contributions of the Protonated and Neutral
Species." Clays Clay Miner., Vol. 35, pp. 121-128.
Anderson, M. A., 1988. "Dissolution of Tetra-
chloroethylene into Ground Water." Ph.D. Disser-
tation, Oregon Graduate Center, Beaverton, OR.
Anderson, M. A., J. F. Pankow, and R. L. Johnson,
1987. "The Dissolution of Residual Dense Non-
aqueous Phase Liquid (DNAPL) from a Saturated
Porous Medium." In: Proceedings, Petroleum
Hydrocarbons and Organic Chemicals in Ground-
water, National Water Well Association and the
American Petroleum Institute, Houston, TX,
November, 1987, pp. 409-428.
Baehr, A. L., 1987. "Selective Transport of
Hydrocarbons in the Unsaturated Zone Due to
Aqueous and Vapor Phase Partitioning." Water
Resources Research, Vol. 23, No. 10, pp. 1926-1938.
Bahr, J. M. and J. Rubin, 1987. "Direct Comparison
of Kinetic and Local Equilibrium Formulation for
Solute Transport Affected by Surface Reactions."
Water Resources Research, Vol. 23, pp. 438-452.
Ball, J. W., D. K. Nordstrom, and E. A- Jenne, 1980.
"Additional and Revised Thermochemical Data
and Computer Code for WATEQ2: A Computerized
Chemical Model for Trace and Major Element
Speciation and Mineral Equilibria of Natural
Waters." U.S. Geological Survey Water Resources
Investigations Nos. 78-116.
Barker, J. F., G. C. Patrick, and D. Major, 1987.
"Natural Attenuation of Aromatic Hydrocarbons
in a Shallow Sand Aquifer." Ground-water
Monitoring Review, Winter 1987, pp. 64-71.
Benjamin, M. M. and J. O. Leckie, 1981. "Multiple-
site Adsorption of Cd, Cu, Zn, and Pb on
Amorphous Iron Oxyhydroxides." J. Coll. Interface
ScL, Vol. 79, No. 2, pp. 209-221.
Chiou, C. T., T. D. Shoup, and P. E. Porter, 1985.
"Mechanistic Roles of Soil Humus and Minerals in
the Sorption of Nonionic Organic Compounds from
Aqueous and Organic Solutions." Organic
Geochemistry, Vol. 8, pp. 9-14.
Chiou, C. T., D. W. Schmedding, and M. Manes,
1982. "Partitioning of Organic Compounds on
Octanol-water Systems." Env. Sci. Tech., Vol. 16,
pp. 4-10.
Chiou C. T., L. J. Peters, and V. H. Freed, 1979. "A
Physical Concept of Soil-water Equilibria for
Nonionic Organic Compounds." Science, Vol. 206,
pp. 831- 832.
Chiou, C. T., P. E. Porter, and D. W. Schmedding,
1983. "Partition Equilibria of Nonionic Organic
Compounds Between Soil Organic Matter and
Water." Env. Sci. Tech., Vol. 17, pp. 227-231.
Coates, John T. and Alan W. Elzerman, 1986.
"Desorption Kinetics for Selected PCB Congeners
. from River Sediment." J. Contaminant Hydrology,
Vol. 1, pp. 191- 210.
Davis, J. A., R. O. James, and J. O. Leckie, 1978.
"Surface lonization and Complexation at the
Oxide/Water Interface: I. Computation of Elec-
trical Double Layer Properties in Simple
Electrolytes." J. Coll. Interface Sci., Vol. 63, No. 3,
480-499.
Dempsey, B. A. and C. R. O'Melia, 1983. "Proton and
Calcium Complexation of Four Fulvic Acid
Fractions." In: Aquatic and Terrestrial Humic
Materials, R. F. Christman and E. T. Gjessing,
Editors. Ann Arbor Science, Ann Arbor, MI.
Drever, J. I., 1982. The Geochemistry of Natural
Waters, Prentice-Hall, Englewood Cliffs, NJ.
Dzombak, D. A., W. Fish, and F. M. M. Morel, 1986.
"Metal-humate Interaction. 1. Discrete Ligand and
Continuous Distribution Models." Environ. Sci.
Technol., Vol. 20, pp. 669-675.
Dzombak, D. A. and F. M. M. Morel, 1986. "Sorption
of Cadmium on Hydrous Ferric Oxide at High
Sorbate/Sorbent Ratios: Equilibrium, Kinetics, and
Modelling." J. Colloid Interface Sci., Vol. 112, No.
2, pp. 588-598.
Dzombak, D. M., 1986. "Towards a Uniform Model
for Sorption of Inorganic Ions on Hydrous Oxides."
Ph.D. Dissertation, Department of Civil Engineer-
ing, Massachusetts Institute of Technology.
53
-------
Ellington, J. J., F. E. Stancil, and W. D. Payne, 1986.
"Measurement of Hydrolysis Rate Constants for
Evaluation of Hazardous Waste Land Disposal:
Volume 1. Data on 32 Chemicals." EPA/600/3-
86/043, Office of Research and Development, U.S.
Environmental Protection Agency, Athens, GA.
Evans, D. W., J. J. Alberts, and R. A. Clark, 1983.
"Reversible Ion-exchange of Cesium-137 Leading
to Mobilization from Reservoir Sediments."
Geochem, Cosmochim. Acta., Vol. 47, No. 11, pp.
1041-1049.
Felmy, A. R., D. C. Girvin, and E. A. Jenne, 1984.
"MINTEQ: A Computer Program for Calculating
Aqueous Geochemical Equilibria." EPA/600/3-84-
032, Environmental Research Laboratory, U.S.
Environmental Protection Agency, Athens, GA.
Fish, W., D. A. Dzombak, and F. M. M. Morel, 1986.
"Metal-humate Interactions. 2. Application and
Comparison of Models." Environ. Sci. Technol.,
Vol. 20, pp. 676-683.
Fu, J. K. and R. G. Luthy, 1986a. "Aromatic
Compound Solubility in Solvent/Water Mixtures."
J. Environ. Eng., Vol. 112, pp. 328-345.
Fu, J. K. and R. G. Luthy, 1986b. "Effect of Organic
Solvent on Sorption of Aromatic Solutes onto
Soils." J. Environ. Eng., Vol. 112, pp. 346-366.
Hanzlik, R. P., 1976. "Inorganic Aspects of Biological
and Organic Chemistry." Academic Press, New
York.
Hayes, M. H. B. and R. S. Swift, 1978. "The Chem-
istry of Soil Organic Colloids." In: The Chemistry of
Soil Constituents, D. J. Greenland and M. H. B.
Hayes, Editors. Wiley Interscience, New York.
Helfferich, F., 1962. Ion Exchange. McGraw-Hill,
New York.
Hem, J. D., 1970. "Study and Interpretation of the
Chemical Characteristics of Natural Water." U.S.
Geological Survey Water-supply Paper 1473.
Honeyman, B. D., 1984. "Cation and Anion
Adsorption at the Oxide/Solution Interface in Sys-
tems Containing Binary Mixtures of Adsorbents:
An Investigation of the Concept of Adsorptive
Additivity." Ph.D. Thesis, Stanford University,
Stanford, CA.
Honeyman, B. D., K. F. Hayes, and J. O. Leckie,
1982. "Aqueous Chemistry of As, B, Cr, Se, and V
with Particular Reference to Fly-ash Transport
Water." Project Report EPRI-910-1, Electric Power
Research Institute, Palo Alto, CA.
Jennings, A. A., D. J. Kirkner, and T. L. Theis, 1982.
"Multicomponent Equilibrium Chemistry in
Ground Water Quality Models." Water Resources
Research, Vol. 18, pp. 1089-1096.
Johnson, R. L., et al., 1985. "Migration of
Chlorophenolic Compounds at the Chemical Waste
Disposal Site at Alkali Lake, Oregon - 2. Contam-
inant Distributions, Transport, and Retardation."
Groundwater, Vol. 23, No. 5., pp. 652-666.
Johnson, R. L., J. A. Cherry, and J. F. Pankow, 1989.
"Diffusive Contaminant Transport in Natural
Clay: A Field Example and Implications for Clay-
lined Waste Disposal Sites." Environ. Sci.
Technol., Vol. 23, pp. 340-349.
Karickhoff, S. W. and K. R. Morris, 1985. "Sorption
Dynamics of Hydrophobic Pollutants in Sediment
Suspensions." Environ. Toxicol. Chem., Vol. 4, pp.
469-479.
Karickhoff, S. W., D. S. Brown, and T. A. Scott, 1979.
"Sorption of Hydrophobic Pollutants on Natural
Sediments." Water Research, Vol. 13, pp. 241-248.
Karickhoff, S. W., 1984. "Organic Pollutant Sorption
in Aquatic Systems." J. Hydraulic Engineer., Vol.
110, pp. 707-735.
Karickhoff, S. W., 1981. "Semi-Empirical Estimation
of Sorption of Hydrophobic Pollutants on Natural
Sediments and Soils." Chemosphere, Vol. 10, pp.
833-846.
Kenaga, E. E. and C. A. I. Goring, 1980. ASTM
Special Technical Publication 707. ASTM,
Washington, DC.
LeBlanc, D. R., et al., 1987. "Fate and Transport of
Contaminants in Septic-contaminated Ground
Water on Cape Cod, Massachusetts." U.S.G.S Open
File Report 87-109, Chapter B.
Lindberg, R. D. and D. D. Runnels, 1984. "Ground
Water Redox Reactions: An Analysis of Equi-
librium State Applied to Eh Measurements and
Geochemical Modeling." Science, Vol. 225, pp. 925-
927.
Mabey, W. R. and T. Mill, 1978. "Critical Review of
Hydrolysis of Organic Compounds in Water Under
Environmental Conditions." J. Phys. Chem. Ref.
Data, Vol. 7, pp. 383-415.
Mabey, W. R., et al., 1982. "Aquatic Fate Process
Data for Organic Priority Pollutants." Chapter 4,
EPA/440/4-81-014, Office of Water Regulations
and Standards, U.S. Environmental Protection
Agency, Washington, DC.
MacKay, D. and B. Powers, 1987. "Sorption of
Hydrophobic Chemicals from Water: A Hypothesis
for the Mechanism of the Particle Concentration
Effect." Chemosphere, Vol. 16, pp. 745-757.
MacKay, D. M., et al., 1986. "A Natural Gradient
Experiment on Solute Transport in a Sand
Aquifer. 1. Approach and Overview of Plume
Movement." Water Resources Research, Vol. 22,
No. 13, pp. 2017-2030.
54
-------
McKay, L. D. and M. R. Trudell, 1987: "Sorption of
Trichloroethylene in Clayey Soils at the Tricil
Waste Disposal Site near Sarnia, Ontario.
Unpublished report, University of Waterloo
Institute for Ground Water Research.
Millington, R. J., 1959. "Gas Diffusion in Porous
Media." Science, Vol. 130, pp. 100-102.
Molz, F. J., et al., 1986. "Performance and Analysis
of Aquifer Tracer Tests with Implications for
Contaminant Transport Modeling." EPA/600/2-
86/062, Office of Research and Development, R.S.
Kerr Environmental Research Lab, U.S. Environ-
mental Protection Agency, Ada, OK.
Morel, F. M. M., 1983. Principles of Aquatic
Chemistry. Wiley Interscience, New York.
Morgan, J. J., 1967. "Application and Limitations of
Chemical Thermodynamics in Water Systems." In:
Equilibrium Concepts in Natural Water Systems,
Advances in Chemistry Series No. 67, American
Chemical Society, Washington, DC.
Murarka, I. P. and D. A. Mclntosh, 1987. "Solid-
waste Environmental Studies (SWES): Descrip-
tion, Status, and Available Results." EPRI EA-
5322-SR, Electric Power Research Institute, Palo
Alto, CA.
Myrand, D., et al., 1989. "Diffusion of Volatile
Organic Compounds in Natural Clay Deposits."
Submitted to J. Contam. Hydrology.
Nkedi-Kizza, P., P.S.C. Rao, and A.G. Hornsby, 1985.
"Influence of Organic Cosolvents on Sorption of
Hydrophobic Organic Chemicals by Soils."
Environ. Sci. Technol., Vol. 19, pp. 975-979.
Palmer, C. D., 1989. "Hydrogeochemistry of the
Subsurface." In: Chemistry of Ground Water, Carl
Palmer, Editor. Lewis Publishers, Chelsea, MI. (In
Review).
Perdue, E. M. and C. R. Lytle, 1983. "A Distribution
Model for Binding of Protons and Metal Ions by
Humic Substances." Environ. Sci. Technol., Vol.
17, pp. 654-660.
Perdue, E. M., 1985. "Acidic Functional Groups of
Humic Substances." In: Humic Substances in Soil,
Sediment and Water, G. R. Aiken, D. M.
MacKnight, R. L. Wershaw, and P. McCarthy,
Editors. Wiley Interscience, New York.
Pye, V. I., R. Patrick, and J. Quarles, 1983. Ground-
water Contamination in the United States,
University of Pennsylvania Press, Philadelphia,
PA.
Rai, D., et al., 1987. "Inorganic and Organic
Constituents in Fossil Fuel Combustion Residues.
Volume 1: A Critical Review." EPRI-EA-5176,
Electrical Power Research Institute, Palo Alto, CA.
Rao, P. S. C., et al., 1985. "Sorption and Transport of
Toxic Organic Substances in Aqueous and Mixed
Solvent Systems." J. Environ. Quality, Vol. 14, pp.
376- 383.
Reichenberg, D. 1966. "Ion Exchange Selectivity."
In: Ion Exchange, Vol. 1., J. A. Marinsky, Editor.
Marcel Dekker, New York.
Roberts, P. V., M. N. Goltz, and D. M. Mackay, 1986.
"A Natural Gradient Experiment on Solute
Transport in a Sand Aquifer: 3. Retardation
Estimates and Mass Balances for Organic Solutes."
Water Resources Research, Vol. 22, No. 13, pp.
2047-2058.
Rueter, J. G., J. J. McCarthy, and E. J. Carpenter,
1979. "The Toxic Effect of Copper on Occillatoria
(Trichodesmium) theibautii." Limnol. Oceanogr.,
Vol. 24, No. 3, pp. 558-561.
Scheigg, H. O., 1984. "Considerations on Water, Oil
and Air in Porous Media." Water Sci. Technol., Vol.
17, pp. 467-476.
Schellenberg, K. C., C. Leuenberger, and R. P.
Schwarzenbach, 1984. "Sorption of Chlorinated
Phenols by Natural Sediments and Aquifer
Materials." Environ. Sci. Technol., Vol. 18, pp.
1360-1367.
Schindler, P. W. and W. Stumm, 1987. "The Surface
Chemistry of Oxides, Hydroxides, and Oxide
Minerals." In: Aquatic Surface Chemistry, W.
Stumm, Editor. Wiley and Sons, New York, pp. 83-
110.
Schindler, P. W., et al., 1976. "Ligand Properties of
Surface Silanol Groups. I. Surface Complex Forma-
tion with Fe3 + , Cu2 + , Cd2 + , and Pb2+." J. Coll.
Interface Sci., Vol. 55, No. 2, pp. 469-475.
Schnitzer, M., 1969. "Reactions Between Fulvic Acid,
A Soil Humic Compound, and Inorganic Soil
Constituents." Soil Sci. Soc. Am. Proc., Vol. 33, pp.
75-81.
Schwarzenbach, R. and J. Westall, 1981. "Transport
of Nonpolar Organic Compounds from Surface
Water to Ground Water: Laboratory Sorption
Studies." Env. Sci. Tech., Vol. 15, pp. 1360-1367.
Schwille, F., 1988. Dense Chlorinated Solvents in
Porous and Fractured Media: Model Experiments.
Translated by J.F. Pankow. Lewis Publishers,
Chelsea, MI.
Sholkovitz, E. R., 1985. "Redox-related Geochem-
istry in Lakes: Alkali Metals, Alkaline Earth
Metals, and Cesium-137." In: Chemical Processes in
Lakes, W. Stumm, Editor. Wiley-Interscience, New
York.
Siegrist, H. and P, L. McCarty, 1987. "Column
Methodologies for Determining Sorption and
55
-------
Biotransformation Potential for Chlorinated Ali-
phatic Compounds in Aquifers." J. Contaminant
Hydrology, Vol. 2, pp. 31-50.
Sillen, L. G., 1967. "Gibbs Phase Rule and Marine
Sediments." In: "Equilibrium Concepts in Natural
Water Systems." Advances in Chemistry Series No.
67, American Chemical Society, Washington, DC.
Smith, R.M. and A.E. Martell, 1976. Critical
Stability Constants. Plenum, New York.
Sposito, G., 1986. "Sorption of Trace Metals by
Humic Materials in Soils and Natural Waters."
CRG Crit. Revs. Environ. Control, Vol. 16, pp. 193-
229.
Sposito, G., 1984. The Surface Chemistry of Soils.
Oxford University Press, New York.
Stevenson, F. J., 1985. "Geochemistry of Soil Humic
Substances." In: Humic Substances in Soil,
Sediment and Water, G. R. Aiken, D. M. McKnight,
R. L. Wershaw, and P. MacCarthy, Editors. Wiley
Interscience, New York.
Stevenson, F. J., 1982 Humus Chemistry: Genesis,
Compositions, Reactions. Wiley Interscience, New
York.
Stumm, W., H. Hohl, and F. Dalang, 1976.
"Interaction of Metal Ions with Hydrous Oxide
Surfaces." Croat. Chem. Acta., Vol. 48, No. 4, pp.
491-504.
Stumm, W. and J. J. Morgan, 1981. Aquatic
Chemistry, 2nd edition. Wiley Interscience, New
York.
Tanford, C., 1973. The Hydrophobic Effect:
Formation of Micelles and Biological Membranes,
Wiley and Sons, Inc., New York.
Thurman, E. M., 1985. "Humic Substances in
Ground Water." In: Humic Substances in Soil,
Sediment and Water, G. R. Aiken, D. M. McKnight,
R L. Wershaw, and P. MacCarthy, Editors. Wiley
Interscience, New York.
Westall, J. C., J. L. Zachary, and F. M. M. Morel,
1976. "MINEQL: A Computer Program for the
Calculation of Chemical Equilibrium Composition
of Aqueous Systems." Technical Note No. 18,
Massachusetts Institute of Technology.
Westall, J. C. and H. Hohl, 1980. "A Comparison of
Electrostatic Models for the Oxide/Solution
Interface." Advances Coll. Interface Sci., Vol. 12,
No. 2, pp. 265-294.
Williams, P. A., 1985. "Secondary Minerals: Natural
Metal Ion Buffers." In: Environmental Inorganic
Chemistry, K. J. Irgolic and A. E. Martell, Editors.
VCH Publishers, Deerfield Beach, FL.
Witkowski, P. J., P. R. Jaffe, and R. A. Ferrara, 1988.
"Sorption arid Desorption Dynamics of Aroclor
1242 to Natural Sediment." J. Contaminant
Hydrology, Vol. 2, pp. 249-269.
Woodburn, K. B., et al., 1986. "Solvaphobic Approach
for Predicting Sorption of Hydrophobic Organic
Chemicals on Synthetic Sorbents and Soils." J.
Contaminant Hydrology, Vol. 1, pp. 227-241.
Wu, S.-c. and P. M. Gschwend, 1986. "Sorption
Kinetics of Hydrophobic Organic Compounds to
Natural Sediments and Soils." Env. Sci. Tech.
Vol. 20, pp. 717- 725.
Zachara, J. M., et al., 1987. "Sorption of Binary
Mixtures of Aromatic Nitrogen Heterocyclic
Compounds on Subsurface Materials." Environ.
Sci. TechnoL, Vol. 21, pp. 397-402.
Zachara, J. M., et al., 1986. "Quinoline Sorption to
Subsurface Materials: Role of pH and Retention of
the Organic Cation." Environ. Sci. TechnoL, Vol.
20, pp. 620-627.
Zachara, J.M., et al., 1988. "Influence of Cosolvents
on Quinoline Sorption by Subsurface Materials
and Clays." J. Contaminant Hydrology, Vol. 2, pp
343- 364.
56
-------
CHAPTERS
SUBSURFACE CHEMICAL PROCESSES: FIELD EXAMPLES
Richard L. Johnson, Carl D. Palmer, and William Fish
Introduction
Understanding chemical processes in the subsurface
is essential for accurate site characterization and for
the design and implementation of efficient remedi-
ation systems. This chapter will focus on examples of
how chemical processes can control transport and
fate in the subsurface.
Petroleum Hydrocarbons in the
Unsatu rated Zone
Gasoline in the subsurface has been examined from
both experimental and theoretical points of view. For
example, theoretical analyses of spills by Baehr
(1987), Corpacioglu and Baehr ;(1987), and Baehr and
Corpacioglu (1987) demonstrate that volatilization is
a dominant mechanism of transport for many
hydrocarbon spills (Figure 58). Gasoline is a complex
mixture of hundreds of compounds, some of which
are more amenable to volatilization than to
dissolution, and vice versa. Also, other components
within gasoline are not particularly prone to either
and will tend to persist in the subsurface.
As discussed in Chapter 5, the importance of
volatilization is determined by the composition of the
free-product, the vapor pressure of the specific
compound, and the rate at which the chemical
diffuses in the vapor phase. For example, aromatic
hydrocarbons in gasoline are moderately volatile,
but also dissolve readily into the pore water.
Therefore, they diffuse more slowly and may persist
in the unsaturated zone longer than the less-soluble
alkanes and alkenes (Baehr, 1987).
Biodegradation also plays an important role in the
fate of gasoline. Many components of gasoline are
REMAINING
IN VADOSE ZONE
50
40
30
20
10
2000
1000
2000
C5 ALKENES
C5-C6 ALKANES
C6 NAPHTHENES
1000
DAYS
2000
Figure 58. (a) Transport pathways for subsurface gaso-
line; (b) masses of aromatics in the
unsaturated zone after a hypothetical spill;
and (c) masses of hydrocarbons in the
unsaturated zone after a hypothetical spill
(after Baehr, 1987).
readily degraded by subsurface microorganisms. In
the saturated zone, biodegradation frequently makes
the aquifer anaerobic, resulting in much slower rates
of degradation. This trend towards anaerobic condi-
57
-------
tions is demonstrated in data collected by Wilson, et
al., (1986) and others at a Traverse City, Michigan
site (Figures 59 and 60).
Heart Of
the Plume
Anaerobic
Zone
Aerobic
Zone
"Renovated"
Zone
Figure 59. Ground-water quality parameters down-
gradient from a hydrocarbon spill at Traverse
City, Michigan (after Wilson, et al., 1986).
In the unsaturated zone, vapor-phase molecular
diffusion can maintain an oxygen supply even at
distances of tens of feet below the ground surface
(Figure 61) (Hult and Grabbe, 1985). The data of
Hult, et al. also show elevated carbon dioxide and
methane concentrations (Figures 61 and 62) which
are the result of biodegradation. In field experi-
ments, Allen, et al. (1987) infiltrated wa'ter
containing simple aromatics into unsaturated porous
media.
Naturally present microorganisms quickly degraded
all of the components. These data, when contrasted
with the fact that the aromatics are widely observed
in the unsaturated zone, suggest that the inter-
actions between gasoline and the subsurface envi-
ronment are complex.
Indicator Compounds
RCRA regulations identify a number of indicator
parameters to be used as an early warning system for
ground-water contamination, including specific
conductance, pH, chloride, bromide, total organic
carbon (TOG), and total organic halides (TOX). In
general, these parameters were selected because
they are conservative and non-reactive, more or less
I
10
8
6
4
2
0
Fe
I
200-
1
-200
eh
Heart of
the Plume
Anaerobic
Zone
Aerobic "Renovated"
Zone zone
Figure 60. Additional ground-water quality parameters
downgradient from a hydrocarbon spill at
Traverse City, Michigan (after Wilson et al.,
1986).
unique to and commonly found in contaminated
ground water, and easily analyzed. Also, compounds
detectable at the leading edge of a contaminant
plume were emphasized.
Most of the RCRA-specified indicator parameters,
however, only partially satisfy the criteria listed
above. Changes in specific conductance, pH, chloride,
and bromide are 'typically not unique to
contaminated groundfwater. Similarly, some natural
total organic carbon .(TOC), chloride, and bromide
are present in virtually all ground water. TOC is
particularly problematic in that natural TOC is often
found in ground water in the range of 1 to 10 mg/L,
whereas it is often desirable to measure organic
contaminants at the ug/L level. Total organic halides
(TOX) determinations have a much lower detection
limit than TOC, but are prone to analytical
difficulties.
Plumb (1985), Plumb and Fitzsimmons (1984), and
Plumb and Pitchford (1985) examined leachate data
from a large number of RCRA and Superfund sites to
determine if other components of leachate would
better serve as indicator parameters. These studies
show that purgeable priority pollutants (PPP) were
found at a high percentage of these sites. The PPP
have low detection limits and are not normally
present in ground water; therefore, they seem an
excellent alternative to the bulk TOC and TOX
58
-------
OXYGIEN (ATM)
CARBON DIOXIDE (ATM)
1420 '
Figure 61. Oxygen and carbon dioxide concentrations in
the unsaturated zone; above an oil spill at
Bemidji, Minnesota (after Hult, et al., 1985).
analyses. However, the sum of l\ll priority pollutants
in landfill leachates often represent only 1 to 5
percent of the total dissolved organic material
present in the ground water (e.g., Reinhard, et al.,
1984; Johnson, et al., 1989'i The bulk-of .the.
unidentified organic contaminants are made up of a
wide variety of primarily hyclrophilic compounds
(Figure 63). An examination of major chemicals
produced in the United Stages (Webber, 1986)
indicates that 7 of the top 35 ape priority pollutants
and only 5 are purgeable priority pollutants (PPP).
Furthermore, many of the tqp 35 chemicals are
significantly more hydrophilic j;han the PPP (Figure
64). This means that if sorption plays a role in
retarding chemical movement, there are many
common compounds that will move more rapidly
than the PPP. Many of those non-PPP compounds
cannot be detected by standard EPA methods.
Nevertheless, they may have (Significant impacts on
ground-water quality due to/«their mobility, solu-
bilityj degradability, toxicity, etc. A better under-
standing of the complexity itff landfill leachate is
necessary before an accurate assessment of the
transport and fate of the wide range of compounds
found there can be made.
TOTAL VOLATILE
HYDROCARBONS (g/m3)
METHANE (PPM)
Figure 62. Hydrocarbon and methane concentrations in
the unsaturated zone above an oil spill in
Bemidji, Minnesota (after Hult, et al., 1985).
Alcohols
Analines
Acetates
Amines
Thiols
Furans
Nitriles
Phenols
Aldehydes
Ketones
Acids
Figure 63. Polar and ionizable compound classes
commonly present in leachates.
Organic Transport on Microparticles
A,growing number of field data suggest that very
hydrophobic compounds can be transported in the
subsurface at rates which are significantly greater
59
-------
Q 15-
U. 13-
§1
§ 9:
Q 7:
g.l
Ul 3
CC .
444-
2,3,4,5,6,2'>5'-
HEPTACHLOROBIPHENYL
NOT PRIORITY OTHER PURGEABLE
PRIORITY POLLUTANTS PRIORITY
POLLUTANTS POLLUTANTS
Figure 64. Estimated retardation factors for common
organics and priority pollutants (foc = 0.01).
than would be predicted by equilibrium sorption
(Means and Wijayaratne, 1982). In addition, nearly
all ground water contains some levels of "dissolved"
organic carbon (DOC) as microparticles. (DOC levels
are commonly 0.1 to 10 mg/L, but may be
significantly higher in the vicinity of landfills or
other waste disposal facilities.)
The importance of microparticle transport of
organics in ground water depends upon the extent of
partitioning to the microparticles. For example,
when the carbon associated with non-settling
particles (NSP) is 1 mg/L, a compound with a Koc of
106 will only experience a decrease in retardation
factor of approximately two. Given the very large
retardation values common for many of these
compounds, a factor of two increase in velocity
generally will not be important. However, if
desorption occurs more slowly than adsorption onto
the microparticles, the relatively rapid movement of
the microparticles could result in significantly
enhanced transport velocities for the hydrophobic
organics. Laboratory data suggest that desorption
times can be quite long, however, the processes
controlling the sorption kinetics of hydrophobic
organics on microparticles currently are not well-
understood.
In a related observation, several investigators
reported that the partition coefficients of hydro-
phobic organics appear to increase at low
sediment/water ratios (Figure 65) (Voice, et al., 1983;
Voice and Weber, 1985; Gschwend and Wu, 1985).
The most satisfactory explanation for this
phenomena is that the process of separating
microparticle-bound organics from the aqueous
phase is incomplete (Gschwend and Wu, 1985). This
effect increases with increasing soil/water ratio, thus
(somewhat counter-intuitively), the Kp values
observed at low soil/water ratios may be more
representative of partitioning in the undisturbed
medium. This observation is important because
107.
(ML/G)
10V
102 1Q3 104
SEDIMENT CONCENTRATION
i(MG/L)
Figure 65. Apparertf: decrease in Kp with increasing
solids concentration in batch equilibrium
tests (adapted from Gschwend and Wu,
1985).
batch tests are frequently used to estimate
retardation factors for many very hydrophobic
compounds at sedilment/water ratios at or above
those in Figure 65. In those cases, Kp values could be
underestimated by ah order of magnitude or more,
leading to a corresponding overestimation of
transport velocity arid an underestimation of the
time required for rertiediation.
Clay Liner Failure
Although clay liners have been a standard method of
containment for land-filled wastes, questions persist
as to their suitability and durability. Three possible
mechanisms of clay-liner failure are advective
transport, waste-induced changes in permeability,
and molecular diffusion. Advective transport
through the liner"can occur if the liner has
mechanically failedy or if a very large hydraulic
gradient exists acro.'ss the liner. An example of the
latter is a landfill filled with water, located in the
unsaturated zone. Under these conditions, gradients
of 10 or higher may make transport velocities
significant, even for at "tight" liner. In such cases,
installation of a gravel sump between two liners can
minimize the hydrautlic gradient across the lower
liner. Advective velocities for a landfill sited in a
saturated medium are expected to be of relatively
60
-------
minor importance, ^except where breaks in the
barrier exist.
There is some experimental evidence that solvents
present in landfills can degrade otherwise
impermeable clay liners (Anderson, et al., 1985;
Brown and Thomas, 1984; Brown and Anderson,
1983; Green, et al., 1983). The main question is
whether or not organic solvents can cause the clay to
shrink and crack. Experiments indicate that if pure
solvents penetrate into the clay, they may be able to
displace water within the clay structure, leading to
shrinkage, cracking, and an increase in perme-
ability. In these experiments, the solvents were
forced into the clay at high pressures equivalent to
many tens of meters of fluid head. This behavior is
consistent with the critical pressure required for
entry of the solvents into the clay (estimated from
the grain diameter of the clay) (Chapter 3; Villaume,
et al., 1983), but is probably not realistic in most
landfill situations.
In the absence of sufficient pressure, entry of the
solvents occurs by simple Fickian diffusion in the
aqueous phase. However, the extent of entry is
limited by the solubilities of the solvent in the water
and vice versa. For many solvents, solubilities are
thousands of milligrams per liter or less, in which
case, the impact of the solvents 6n the clay is
expected to be minimal.
The presence of mechanical failures or other
imperfections in the clay may provide focal points for
degradation of the clay liners. Miscible liquids, such
as alcohols, acids, bases, and ketones, also have the
potential to affect clay structure. Because they are
not limited by solubility, large concentrations can
diffuse into the clay. Miscible compounds generally
are polar molecules which can interfere with the
electrostatic forces within the clay (Fernandez and
Quigley, 1985). This interference can result in a
breakdown of the clay structure and subsequent
failure of the liner.
Even though chemical concentrations of thousands of
milligrams per liter within clay liners are not
expected to affect the hydraulic properties of the
clay, such concentrations could have a major impact
on ground-water quality because of the diffusive
movement of chemicals through the liner. In the
absence of advection, mass transport is controlled by
Pick's Second Law:
\ a2c
R
dt
(2)
A number of field studies demonstrated the
importance of simple Fickian diffusion (Johnson, et
al., 1989; Goodall and Quigley, 1977; Crooks and
Quigley, 1984; Desaulniers, et al., 1981). Johnson, et
al. collected clay cores from beneath a five-year-old
waste disposal cell and determined concentration
profiles for chloride and several organic contam-
inants. The chloride diffused nearly one meter while
the organics moved much shorter distances. The
principal reason for the slower diffusion of the
organics was their sorption onto the clay (foc~0.01).
For non-sorbed contaminants (including organics
when the foc is low), breakthrough of a one-meter
liner could occur in less than 10 years. Within a few
decades, steady-state concentrations could develop
across the clay, possibly resulting in significant mass
transfer into the underlying aquifer.
Under steady-state conditions, mass flux follows
Pick's First Law:
ac
D9 —
e
(3)
As the calculation in Figure 66 suggests, the mass
flux through a liner via diffusion can be on the order
of 1 gram per square meter per year. For a large
liner, a mass transfer of thousands of grams of
individual contaminants per year to the aquifer
1000 ppm
1m
0 ppm
dC
t
= -0.37 x 1.5x10"6 cm2 /sec x 1g/L-m
D
e ,2
dX
ac
dt
(1)
As seen in Chapter 5, for sorbing compounds, Fick's
Law can be modified to handle equilibrium
partitioning:
= 5.55x10 g/cm /sec
•» 1.75 g/m /year
Figure 66. Example calculations of steady-state diffu-
sive mass flux through a one-meter thick clay
liner.
61
-------
could occur. Similar conditions can exist across
natural aquitards as well, and the large contact area
between aquifers could result in significant mass
transfer into an uncontaminated aquifer.
"Plume Sniffing"
Because soil vapor samples are relatively easy to
collect from the subsurface, interest has grown in
using soil-gas sampling or "plume sniffing" to search
for ground-water contamination (Marrin and
Kerfoot, 1988). Shallow soil-gas sampling detects
vapors emitted into the unsaturated zone by ground-
water contaminant plumes that contain volatile
organics. Many factors affect vapor concentrations in
the unsaturated zone. For example, vertical trans-
verse dispersion in the saturated zone may control
mass flux into the unsaturated zone. Because air-
diffusion coefficients are much greater than the
values typically reported for vertical dispersion (10-2
cm2/s vs. 10'6 cm2/s), unsaturated zone vapor concen-
trations might be expected to be low. Nevertheless,
substantial unsaturated concentrations have been
reported at several sites. Swallow and Gschwend
(1984) examined the steady-state distribution of
volatile organic vapors in the ground water and
unsaturated zones (Figure 67). For vertical trans-
verse dispersivities (atv) in the ground water of 1 cm
or less, concentrations at steady-state were found to
be near zero through the unsaturated zone. A
limitation of the steady-state approach is that data
are not provided during the time required for the
system to reach steady-state. For the 10-meter
aquifer in the Swallow and Gschwend model system,
with atv = 1 cm, hundreds of years would be required
before the system reached the steady-state.
As previously mentioned, vapor concentrations in
the unsaturated zone may deviate from idealized
ui
ui
UNSATURATED
CAPILLARY
SATURATED
CONCENTRATION
Figure 67. Schematic drawing of Steady-state concen-
tration profiles for several vertical dispersion
values (adapted from Swallow and
Gschwend, 1984).
values due to environmental conditions or biodegra-
dation. Data from Hult and Grabbe (1985) (Figure
68) for the Bemidji, Minnesota site show that vapor
concentrations at the depths typically used for plume
sniffing (3 to 6 feet) are very low because of
biodegradation. This is the case even though a large
organics source is present as free-product at the
capillary fringe. Vapors from ground-water plumes
also can be masked by clean water which forms a cap
over the plume. This occurs if local recharge
displaces the contaminants from the capillary fringe,
or if the plume moves downward into the ground
water.
Numerous case histories demonstrate the usefulness
of plume sniffing (Marrin and Thompson, 1987;
Marrin and Kerfoot, 1988), especially for chlorinated
solvents. As a result, plume sniffing continues to
receive a great deal of attention, despite the fact that
its general applicability has not been thoroughly
demonstrated. As Marrin and Kerfoot (1988)
correctly point out:
Conventional technologies available for sub-
surface investigation (e.g., monitoring .wells
and soil borings) always will be required to
confirm and monitor subsurface contamina-
tion; however, quicker and less expensive tech-
niques are useful for preliminary site evalu-
ations.
They also suggest that, while positive results
generally suggest detection of ground-water plumes,
failure to detect plumes does not assure that ground-
water contamination is not present.
Ground-Water Contamination by
Chromium
Chromium (Cr) in the environment causes great
concern because of its wide use in industry and its
potentially high toxieity. Although chromium can
exist in oxidation states ranging from -2 to + 6, the
trivalent (Cr(III)) and hexavalent (Cr(VT)) species
are the common stable forms found in the environ-
ment. Three recent surveys of the environmental
significance of chromium (U.S. EPA, 1983; Radian
Corp., 1983; the American Petroleum Institute,
1981) emphasize that the chemical form of this metal
determines its environmental behayior and toxieity.
Cr(III) is relatively insoluble and exhibits little or no
toxieity (van Weerelt, et al., 1984), while Cr(VI)
usually occurs as the highly soluble CrO42-, HCrO^,
and Cr2O72- anions.
Cr(III) is the most common form of chromium in the
earth's crust; the predominant source of hexavalent
chromium in the environment is anthropogenic
activities. For example, industry has used Cr(VI)
62
-------
TOTAL VOLATILE HYDROCARBONS
UJ
O
<
1L
DC
O
Hi
UJ
LU
IL
30
RELATIVE CONCENTRATION
Figure 68.
Vertical hydrocarbon profiles in the unsaturated zone above an oil spill near Bemidji, Minnesota site (adapted
from Hult, et al., 1985) .
primarily in metal plating and leather tanning
applications for over 100 years. As a result,
numerous waste lagoons, dumps, and landfills are
contaminated with chromate wastes (Black and Heil,
1982; Cook and DiNitto, 1982; Owen, 1982; Massa-
chusetts Department of Environmental Quality
Engineering, 1981; Keely and Boateng, 1987). At
many sites, the chromium coexists with a variety of
other inorganic and organic wastes, and under these
conditions, a wide range of chemical interactions are
possible. t
Chromium transport in aqueous systems strongly
depends on sorption, chelation, and redox reactions.
The redox reactions are only poorly understood, yet,
they are of key importance because the redox state of
chromium dictates its sorptive and chelation
behavior. Despite the strong tendency for chromium
to partition to many mineral surfaces, there are
reports of extensive chromium migration (Keely and
Boateng, 1987).
Surface sorption is an important element of
chromium behavior. Electrostatic forces near the
solid-liquid interface form a transition range near
neutral pH in which both positively and negatively
charged surface sites coexist (Swallow, et al., 1980).
Several published studies discuss adsorption of
chromium species at oxide interfaces (Bartlett and
Kimble, 1976; Bartlett and James, 1979; Huang and
Wu, 1977). If the mineral composition of the aquifer
is known, appropriate sorption parameters for Cr(III)
or Cr(VI) can be roughly estimated. Unfortunately,
the redox state of the chromium also must be known
and this is difficult to predict because natural redox
chemistry of Cr(HI)/Cr(VI) is not well understood.
Although Cr(III) and Cr(VT) are each quite stable
and tend to be kinetically inhibited from undergoing
redox transformations, there are systems which
catalyze oxidation-reduction reactions of chromium.
In strong acid solutions, Cr(VI) will oxidize organic
compounds and be reduced to Cr(III) (Bartlett and
Kimble, 1976). This reaction eliminates toxic Cr(VI)
and generates the relatively insoluble trivalent
species. The reverse reaction can be driven by MnC>2
in a process in which the MnOg appears to act as both
an oxidizing agent and a catalytic surface (Bartlett
63
-------
and James, 1979). James and Bartlett (1983)
reported the oxidation of Cr(III) in a mixture of
tannery sludge and moist soil containing MnC>2.
Thus, Cr(VI) may be generated at waste sites in
which only Cr(III) is deposited. It is possible that at
an actual waste disposal site, the organic oxidation
and Mn02 reduction reactions may form a cycle.
Another chromium/organic association that is
largely unexplored yet potentially important to
chromium behavior at waste sites is the solvation of
chromate by organic phases. Essentially all studies
of chromium in the environment focus on the solid-
or aqueous-phase chromium. However, chromate is
typical of a class of relatively large inorganic anions
that dissolve to an appreciable extent in certain
organic solvents. This phenomenon is well docu-
mented for highly acidic conditions in analytical
solvent extraction studies (Zolotov, et al., 1967).
Recent studies show that organic solvation of
chromate occurs even near neutral pH; the phe-
nomenon, therefore, has wider environmental
significance than would be deduced from the solvent
extraction literature. Reactions also may be
enhanced by the close association of the chromate
with the solvent.
Partitioning of chromate into non-aqueous phases
may actually increase the mobility of chromium in
the environment. Both Cr(VI) and dense chlorinated
solvents are frequently associated with plating
operations. This suggests that the chromate could
become stabilized in the dense solvents and be
transported well below the water table.
In summary, reactions between solvated Cr(VI) and
organic compounds may affect the composition and
behavior of ground-water contaminants by: (1)
removing more hexavalent chromium from the
aqueous phase than would occur by simple solvation;
(2) stabilizing Cr(VI) in organic phases; and (3)
mediating the oxidation of organic waste components
into new compounds with greater or lesser toxicity
and mobility.
Ground-Water Contamination by Leaded
Gasoline
The subsurface movement of lead is expected to be
minimal in most aquifers because of its low solubility
and strong tendency to sorb to aquifer materials.
However, studies have found that gasoline-derived
lead can be transported over hundreds of meters. The
mechanism(s) of this transport remain unclear,
although transport as the alkyl lead, on
microparticles, and/or in the free-product gasoline
may be involved. An example case history of lead
movement was conducted at the site of a large
gasoline spill in Yakima, Washington (Fish, 1987).
Up to 20,000 gallons of leaded and unleaded gasoline
were released over several years. At the time of
sampling, the lead had been transported over 100
meters in a heterogeneous alluvial aquifer.
At least three hypotheses have been made to explain
these data. The first is that some of the lead is
associated with fine-colloidal particles transported
through the aquifer. The presence of large quantities
of lead in the solid phase would support this possi-
bility. However, the relatively low concentrations
observed in ground water filtered through a 0.1 um
filter suggests that if particle transport is the mecha-
nism, the bulk of the particles must be greater than
0.1p.m.
A second hypothesis is that the bulk of the lead
transport occurred when the lead existed in less
strongly sorbed alkylated forms. However, no alkyl-
ated lead could be found in the aquifer at the time of
sampling.
A third hypothesis is that the bulk of the transport of
the lead, as well as the organics, took place while
both were associated with the free-product gasoline.
The volume of the spill is such that the product could
have extended over significant distances. However,
as discussed in Chapter 3, the erratic behavior of
immiscible phases in the subsurface is poorly under-
stood. At the time of sampling, free product could be
observed in wells at distances of up to 100 m from the
source. The highly heterogenous, interbedded nature
of the aquifer also might allow relatively narrow,
preferential pathways for free-product movement
that could enhance traveling distances.
References
Allen, R. M., R. W. Gillham, and J. F. Barker, 1987.
"Remediation of Gasoline-contaminated Ground
Water by Infiltration Through Soil." National
Water Well Association, Fourth Annual Eastern
Regional Ground Water Conference, July 14-16,
1987, Burlington, VT.
American Petroleum Institute (API), 1981. Sources,
Chemistry, Fate, and Effects of Chromium in
Aquatic Environments, American Petroleum
Institute. ,
Anderson, D. C., K. W. Brown, and J. C. Thomas,
1985. "Conductivity of Compacted Clay Soils to
Water and Organic Liquids." Waste Manage. Res.,
Vol. 3, No. 4, pp. 339-349.
Baedecker, M. J., et al., 1985. "Geochemistry of a
Shallow Aquifer Contaminated with Creosote
Products." In: Proceedings, Second U.S. Geological
Survey Toxic Waste Technical Meeting, Cape Cod,
MA, October, 1985.
64
-------
Baehr, A. L. and M. Y. Corpacioglu, 1987. "A
Compositional Multiphase Model for Ground-water
Contamination by Petroleum Products. 2:
Numerical Solution." Water Resources Research,
Vol. 23, No. 1, pp. 201-213.
Baehr, A. L., 1987. "Selective Transport of Hydro-
carbons in the Unsaturated Zone Due to Aqueous
and Vapor Phase Partitioning.1' Water Resources
Research, Vol. 23, No. 10, pp. 1926-1938.
Bartlett, R. and B. James, 1975). "Behavior of
Chromium in Soils: III. Oxida.tion." J. Env.
Quality, Vol. 8, pp. 31-35.
Bartlett, R. and J. M. Kimble, 197:6. "Behavior of
Chromium in Soils: II. Hexavalerit Forms." J. Env.
Quality, Vol. 5, pp. 383-388.
Black, J. A. and J. H. Heil, 1982. "Municipal Solid
Waste Leachate and Scavenger Waste: Problems
arid Prospects in Brookhaven Town." In:
Proceedings, N.E. Conference on the Impact of
Waste Storage and Disposal on Ground-water
Resources, R. P. Novitsky and G. Levine, Editors.
U.S. Geological Survey/Cornell University.
Brown, K. W. and D. C. Anderson,JL983. "Effects of
Organic Solvents on the PermeaTbility of Clay
Soils." EPA/600/2-83-016, Office of Research and
Development, U.S. Environmental Protection
Agency. .T,.
Brown, K. W. and J. C. Thomas, 1984. "Conductivity
of Three Commercially Available Clays to Petrol-
eum Products and Organic Solvents." Hazard.
Waste, Vol. 1, No. 4, pp. 545-553.
Cook, D. K. and R. G. DiNitto, 1982. '"Evaluation of
Ground Water Quality in East and'North Woburn,
MA." In: Proceedings, N.E. Conference on the
Impact of Waste Storage and Disposal on Ground-
Water Resources, R. P. Novitsky and G. Levine,
Editors. U.S. Geological Survey/Cornell Uni-
versity. ;
Corpacioglu, M. Y. and A. L. Bae.hr, 1987. "A
Compositional Multiphase Model for Ground-water
Contamination by Petroleum Products. 1: Theo-
retical Considerations." Water Resources Research,
Vol. 23, No. 1, pp. 191-200.
Crooks, V. E. and R. M. Quigley, 1984. "Saline
Leachate Migration through Clay: A Comparative
Laboratory and Field Investigation." Can. Geotech.
J., Vol. 21, pp. 349-362.
Desaulniers, D. E., J. A. Cherry, and P. Fritz, 1981.
"Origin, Age, and Movement of Pore Water in
Argillaceous Quaternary Deposits at-JFour Sites in
Southwestern Ontario." J. Hydrology, Vol. 50, pp.
231-257.
Fernandez, F. and R. M. Quigley, 1985. "Hydraulic
Conductivity of Natural Clays Permeated with
Simple Liquid Hydrocarbons." Can. Geotech. J.,
Vol. 22, No. 2, pp. 205-214.
Fish, W., 1987. "Subsurface Transport of Gasoline-
derived Lead at a Filling Station Contamination
Site in Yakima, Washington." In: Proceedings,
NWWA FOCUS: Northwestern Ground Water
Issues Conference, National Water Well Associ-
ation, Portland, OR, May 1987.
Godsy E. M. and D..F. Goerlitz, 1984. "Anaerobic
Microbial Transformations of Phenolic and Other
Selected Compounds in Contaminated Ground
Water at a Creosote Works, Pensacola, FL." In:
Movement and Fate of Creosote Waste in Ground
Water, Pensacola, FL, U.S. Geological Survey Open
File Report 84-466.
Goodall, D. C. and R. M. Quigley, 1977. "Pollutant
Migration from Two Sanitary Landfill Sites near
Sarnia, Ontario." Can. Geotech. J., Vol. 14, pp. 223-
236.
Green, W. J., et al., 1983. "Interaction of Clay Soils
with Water and Organic Solvents: Implications for
the Disposal of Hazardous Wastes." Environ. Sci.
TechnoL, Vol. 17, No. 5, pp. 278-282.
Gschwend, P. M. and S.-c. Wu., 1985. "On the
Constancy of Sediment-water Partition Coeffi-
cients of Hydrophobic Organic Pollutants." Env.
Sci. Tech., Vol. 19, pp. 90-96.
Huang, C. P. and M. H. Wu, 1977. "The Removal of
Chromium (VI) from Dilute Aqueous Solution by
Activated Carbon." Water Res., Vol. 11, pp. 673-
678.
Hult, M. F. and R. R. Grabbe, 1985. "Permanent
Gases and Hydrocarbon Vapors in the Unsaturated
Zone." In: Proceedings, U.S. Geological Survey
Second Toxic-Waste Technical Meeting, Cape Cod,
MA, October, 1985.
James, B. R. and R. J. Bartlett, 1983. "Behavior of
Chromium in Soils: IV. Interactions Between
Oxidation-reduction and Organic Complexation."
J. Env. Quality, Vol. 12, pp. 173-176.
Johnson, R. L., C. D. Palmer, and J. F. Keely, 1987.
"Mass Transfer of Organics Between Soil, Water,
and Vapor Phases: Implications for Monitoring,
Biodegradation, and Remediation." In:
Proceedings, Petroleum Hydrocarbons and Organic
Chemicals in Ground Water, National Water Well
Association and the American Petroleum Institute,
Houston, TX, November, 1987.
Johnson, R. L., J. A. Cherry, and J. F. Pankow, 1989.
"Diffusive Contaminant Transport in Natural
Clay: A Field Example and Implications for Clay-
lined Waste Disposal Sites." Accepted for
publication in Environ. Sci. TechnoL
65
-------
Keely, J. F. and K. Boateng, 1987. "Monitoring Well
Installation, Purging, and Sampling Techniques.
Part 2: Case Histories." Groundwater, Vol. 25, No.
4, pp. 427-439.
Marrin, D. L. and G. M. Thompson, 1987. "Gaseous
Behavior of TCE Overlying a Contaminated
Aquifer." Groundwater, Vol. 25, pp. 21-27.
Marrin, D. L. and H. B. Kerfoot, 1988. "Soil-gas
Surveying Techniques." Environ. Sci. Technol.,
Vol. 22, No. 7, pp. 740-745.
Massachusetts Department of Environmental
Quality Engineering (DEQE), 1981. "Interim
Report on Chemical Contamination in Massa-
chusetts."
Mattraw, H. C. and B. J. Franks, 1984. "Description
of Hazardous Waste Research at a Creosote Works,
Pensacola, FL." In: Movement and Fate of Creosote
Waste in Ground Water, Pensacola FL, U.S.
Geological Survey Open-File Report, pp. 84-466.
Means, J. C. and R. Wijayaratne, 1982. "Role of
Natural Colloids in the Transport of Hydrophobic
Pollutants." Science, Vol. 215, pp. 968-970.
Owen, K. C. 1982. "Ground-water Resources in
Niagara Falls, NY and the Potential Impacts of
Hazardous Waste Contamination." In: Proceed-
ings, N.E. Conference on the Impact of Waste
Storage and Disposal on Ground-water Resources,
R. P. Novitsky and G. Levine, Editors. U.S.
Geological Survey/Cornell University.
Plumb, R. H., 1985. "Disposal Site Monitoring Data:
Observations and Strategy Implications." In:
Proceedings, Second Annual Canadian/American
Conference on Hydrogeology: Hazardous Wastes in
Ground Water — A Soluble Dilemma, National
Water Well Association, Dublin, OH, June, 1985.
Plumb R. H. and C. K. Fitzsimmons, 1984.
"Performance Evaluations of RCRA Indicator
Parameters." In: Proceedings, Second Annual
Canadian/American Conference on Hydrogeology:
Practical Applications of Geochemistry, National
Water Well Association, Dublin, OH, June, 1985,
pp. 125-137.
Plumb, R. H. and A. M. Pitchford, 1985. "Volatile
Organic Scans: Implications for Ground-water
Monitoring." In: Proceedings, Petroleum Hydro-
carbons and Organic Chemicals in Ground Water,
National Water Well Association and the Amer-
ican Petroleum Institute, Houston, TX, November,
1985, pp. 13-15.
Radian Corp., 1983. Estimates of Population Expo-
sure to Ambient Chromium Emissions, Radian
Corp., Durham, NC.
Reinhard, M., N. L. Goodman, and J. F. Barker,
1984. "Occurrence and Distribution of Organic
Chemicals in Two Landfill Leachate Plumes."
Environ. Sci. Technol., Vol. 18, No. 12, pp. 953-961.
Swallow, J. A. and P. M. Gschwend, 1984. "Volatil-
ization of Organic Compounds from Unconfined
Aquifers." In: Proceedings, Petroleum Hydro-
carbons and Organic Chemicals in Ground Water,
National Water Well Association and the
American Petroleum Institute, Houston, TX,
November 1984.,
Swallow, K. C., D. N. Hume, and F. M. M. Morel,
1980. "Sorption, of Copper and Lead by Hydrous
Ferric Oxide." Environ. Sci. Technol., Vol. 14, No.
8, pp. 1326-13311..
U.S. Environmental Protection Agency, 1983.
"Health Assessment Document for Chrome." EPA-
600/83-104A.
van Weerelt, M., W. C. Pfeiffer, and M. Fiszman,
1984. "Uptake and Release of Chromium (VI) and
Chromium (HI) by Barnacles." Mar. Environ. Res.,
Vol. 11, No. 3, pp. 201-211.
Villaume, J. F., IP: C. Lowe, and D. F. Unites, 1983.
"Recovery of' Coal Gasification Wastes: An
Innovative Approach." In: Proceedings, Third
National Symposium on Aquifer Restoration and
Ground-water ^Monitoring, National Water Well
Association, V/orthington, OH, pp. 434-445.
Voice, T. C., C.rJp; Rice and W. J. Weber, Jr., 1983.
"Effect of Solids Concentration on the Sorptive
Partitioning of Hydrophobic Pollutants in Aquatic
Systems." Em*-. Sci. Technol., Vol 17, pp. 513-518.
Voice, T. C. anjd W. J. Weber, Jr., 1985. "Sorbent
Concentration! Effects in Liquid/Solid Partition-
ing." Env. Sci, Technol., Vol 19, pp. 789-796.
Webber, D., 19;86. "Top 50 Chemicals Production
Dropped Modlerately in 1985." Chem. Eng. New,
April 21,198B,-pp. 13-15.
Wilson, B. H./ et al., 1986. "Biological Fate of
Hydrocarboris at an Aviation Gasoline Spill Site."
In: Proceedings, Petroleum Hydrocarbons and
Organic Chemicals in Ground Water, National
Water Well Association and the American Petrol-
eum Institute, Houston, TX, November, 1986.
Zolotov, Y. A.,[V. V. Bagreev, and O. M. Petrukhin,
1967. "Solvent Extraction Chemistry." In: Pro-
ceedings, Inil. Conference on Solvent Extraction
Chemistry, Gothenburg, North-Holland, Amster-
dam. '• '•
66
-------
CHAPTER?
MICROBIAL ECOLOGY AND POLLUTANT BIODEGRADATION IN
SUBSURFACE ECOSYSTEMS
Joseph M. Suflita
Introduction
Interest in ground water-related sciences is growing,
and individuals involved with all facets of ground^
water use and protection are developing a greater
awareness of the chemistry, physics, hydrology,
geology, and biology of the terrestrial subsurface.
Information gathered from ground-water contam-
ination sites helps to focus questions regarding the
possibility that aquifers possess natural attenuation
mechanisms for pollutant abatement. Consideration
of this possibility has led investigators to realize that
natural physico-chemical processes may, at best,
only partially transform aquifer contaminants.
Parallel to the general interest in ground water came
the beginnings of serious study of ground-water
microbiology. Because of the biochemical versatility
of natural microbial communities, investigators
thought that the microflora indigenous to aquifers
might function to remove problem contaminants. In
fact, microbial metabolism of ground-water pollu-
tants is the only technology that has the potential to
completely degrade pollutants in situ and convert
them to more environmentally acceptable forms.
This chapter will introduce the historic and current
scientific perspectives regarding microbial ecology of
the terrestrial subsurface and will focus on how these
perceptions evolved. Examples will be given of the
diverse types of subsurface microorganisms, micro-
bial communities, and associated metabolic activ-
ities. Also, the metabolic principles governing pollu-
tant biodegradation in other habitats will .be
extrapolated to ground water. Limits to pollutant
biodegradation will be considered in the context of
existing environmental conditions, physiology of the
indigenous microflora, and chemical structure of the
contaminants. Finally, this chapter will discuss how
these principles can be applied to either the in situ or
above-ground bioremediation of contaminated
aquifers.
Historical Perceptions
Several historical misconceptions about ground-
water microbiology hindered the use of bio-
reclamation and probably contributed to the abuse of
aquifers. First, ground water was perceived as a safe
water source, protected by a metabolically diverse
"living filter" of microorganisms in the soil root zone
which functioned to convert organic pollutants to
innocuous endproducts. However, recent evidence
indicates that, in many instances, ground water may
be at least as contaminated as surface waters
(Council on Environmental Quality, 1981; Page,
1981). In retrospect, the historical perspective is
somewhat understandable since ground-water flow
is generally slow (10 to 100 m per year) and transport
processes are complex (McKay, et al, 1985). A
significant time lag for the movement of chemicals
from their subsurface source to even nearby wells is
common, and many years can pass before environ-
mental or health impacts of ground-water contam-
ination become evident.
Secondly, aquifers were considered abiotic environ-
ments since early studies indicated that microbial
numbers decreased with soil depth (Waksman,
1916), and later studies showed that the vast
majority of cells were attached to soil particulates
(Balkwill, et al., 1977). In addition, by estimating the
time required for surface water to vertically
penetrate subsurface formations, investigators
reasoned that microbes traveling with the water
67
-------
would quickly exhaust available nutrients and
rapidly die off. These beliefs led to the perception
that aquifers were sterile and essentially irre-
trievably lost when contaminated by polluting chem-
icals. Now, however, microscopic, cultivation, meta-
bolic, and biochemical investigations, often using
aseptically obtained aquifer material, reveal that the
terrestrial subsurface harbors a surprisingly rich
assemblage of procaryotic and some eucaryotic life
forms. The evidence supporting this perception is
described in the following sections.
Ground-Water Contamination
The sources of ground-water contamination have
important implications for the design of biorestora-
tion techniques. Contaminants can originate from
non-point sources, such as agricultural chemicals
and road salts, and from point sources including
residential septic systems, leaking underground
storage tanks, surface impoundments, landfills, and
transportation losses. Comprehensive reviews of the
sources and types of ground-water pollution are
presented by Keswick (1984); Craun (1984); and
Zoetman (1985).
It is important to point out that ground-water
pollution is not just a water pollution problem. As
shown in Figure 69, a spill of gasoline hydrocarbons
tends to exist in multiple phases. The gasoline free
product represents the most severe form of contam-
ination, but also is the most limited in terms of area
affected. As gasoline moves from a spill site, it
contaminates soil in the vadose zone. Since its
components are largely water-insoluble and less
dense than water, hydrocarbon free product tends to
reside and spread along the water table boundary.
Hydrocarbon free product can easily pollute wells
within the zone of contamination, and also can sorb
to and contaminate those soil areas influenced by
water table fluctuations. This sorbed material tends
to be another more subtle source of secondary
contamination. In addition, free product can contam-
inate surface waters, and gasoline vapors can collect
in basements of buildings and create inhalation or
explosion risks.
As discussed in earlier chapters, the severity of soil
contamination tends to be a function of the proper-
ties of the soil. Water, too, affects the amount of
contamination because it is an excellent solvent and
can even dissolve, to some extent, those substances
deemed insoluble. The concentrations of hydrocar-
bons in the water phase are generally low, but the
amount of area affected by such contamination can
be very high. Consequently, this form of contamina-
tion has the highest potential for human exposure.
All remediation technologies must consider where
the contaminant resides in the subsurface; treatment
of only the water phase may provide a temporary
solution to an immediate problem, but also may
effectively ignore the more subtle and, many times,
more important problem areas.
Evidence for Subsurface Microbiota
Microbiological investigations of the terrestrial
subsurface have revealed that all aquifers examined
thus far possess a sometimes surprisingly rich micro-
flora. Such studies include direct microscopic, culti-
vation, metabolic, and biochemical evidence for
microorganisms in aseptically obtained aquifer
material. Table 4 is only a partial listing of the
typical numbers of bacteria detected in various
geological settings. This table illustrates that rela-
tively high numbers of microorganisms can be
detected in both contaminated and pristine aquifers
of varying depth and geological composition. There is
Contaminated
Water
•Supply
I Fume/ Explosion
Hazard
site
Surface
Contamination
Contaminated
Soil
Water
Table
Free Product
Contaminated Water
Figure 69. Multiple problems associated with hydrocarbon release to the terrestrial subsurface.
68
-------
Table 4. Microbial Numbers in Various Geological Settings and Depths
Aquifer Site
Sand/Gravel
Lula, OK
Sand
Pickett, OK
Loamy Clay
Fort Polk, LA
Sand/Clay
Conroe, TX ,
Burn Cavern
Tenn. Colony, TX
Carrizo Aquifer
TX
Mogothy Aquifer
Long Island, NY
Sand
Ontario, Canada
Gravel
Dayton, OH
Glacial Till
St Louis Pk., MN
Marmot Basin
Alberta, Canada
Bacatunna, Clay
Pensacola, FL
Sand/Gravel
Gape Cod, MA
Sand
Norman, OK
Sample
Type3
S
S
S
S
S
w
w
w
w
w
w
w
S
S
S
Depth
(m)
5
5.5
5
5.5
7.5
91
76
146
10
10-12
25-35
1.5
410
12-32
1.8
Aquifer
Contaminant
None
None
None
None
Creosote
Phenolics
None
Treated sewage
Septage
None
Creosote
None
Acid waste
Sewage
Landfill leachate
Total Countb
x 106
Cells/gdw
or Cells/ml
3.8-9.3
5.2
9.8
1-1 0<*
5-49
—
™
-
0.14
0.036-0.06
0.07-10
0.05-2.5
10<*
11-34
11-17
Viable Count<=
x 104
CFU/gdw
or Cells
20-110
250
-
—
4.2
0.04
0.01-5
0.031
0.052-0.06
0.09-4006
*"
-
~
~
References
Wilson, etal., 1983;
Balkwill and Ghiorse, 1985
Balkwill and Ghiorse,
1 985; Ghiorse and
Balkwill, 1985
Ghiorse and Balkwill,
1983; White, etal., 1985
Webster, et al., 1 985
Humenick, etal., 1982
Humenick, et al., 1 982
Godsey and Ehrlich, 1978
Ventullo and Larson, 1 985
Ventullo and Larson, 1 985
Ehrlich, etal., 1983
Ladd, etal., 1982
White, etal., 1983
Harvey, et al., 1 984
Beeman and Suflita, 1 987
aAquifer Solids
"Ground Water
*>Acridine Orange Direct Count (Unless Otherwise Noted)
cPlate Count Assay (Unless Otherwise Noted)
dSignature Lipid Analysis
eMost Probable Number Assay
good evidence that even deep geological formations
can be suitable habitats for bacteria (Kuznetsov, et
al., 1963; Updegraff, 1982).
The picture that emerges from microbiological
studies is that subsurface microorganisms tend to be
small, capable of response to the influx of nutrients,
and primarily attached to solid surfaces. Although
detectable, eucaryotic microorganisms are few in
number relative to surface soil microorganisms and
presumably of minor importance in the terrestrial
subsurface. Microeucaryotes (such as fungi and
protozoa) probably exist in most aquifers as
metabolically inert resting structures. These data
reinforce information published in a recent book
(Bitton and Gerba, 1984) which reiterates that there
is little to preclude microbial growth in the terres-
trial subsurface (McNabb and Dunlap, 1975). This
conclusion is important because it shows that those
aquifers that are most susceptible to contamination
by chemicals (those less than 100 meters deep), do
possess diverse microbial communities.
Sample Procurement
The development of suitable sampling technology
has helped to show the existence of a sizable sub-
surface microbiota. Any microbiological probe of the
terrestrial subsurface is critically dependent on the
quality, integrity, and representativeness of the
samples obtained. To make valid interpretations of
69
-------
the data resulting from subsurface investigations,
the samples received must not be contaminated with
nonindigenous microorganisms. Potential sources of
contamination include organisms present on the
drilling machinery, in surface soil layers, in drilling
muds, and in water used to make up the drilling
muds. Since most subsurface biology is associated
with the solid matrix, the discussion of sampling will
be limited to aquifer sediments. However, McNabb
and Mallard (1984) have published an excellent
review of all the requirements for microbiological
sampling of aquifers.
No sampling effort can be performed without some
microbiological contamination, regardless of
whether drilling tools are employed to reach deeper
regions of the subsurface or if excavation pits are dug
in relatively shallow areas. With this realization,
most current sampling efforts rely on the recovery
and subsequent dissection of cores to remove
microbiologically compromised portions. This dissec-
tion can be performed in the field or, if the sampling
sites are nearby, when the cores are returned to the
laboratory. In all cases, the outer few centimeters
and the top and bottom portions of aquifer cores are
removed because of possible contamination by nonin-
digenous bacteria, and only the center portions of an
aquifer core are used for subsequent analysis.
Ideally, this dissection process occurs as soon as
possible after the core is removed from the ground so
that nonindigenous microorganisms do not have a
chance to penetrate to the inner portions of the core.
In the field, an alcohol-sterilized paring device is
used in the dissection process (McNabb and Mallard,
1984). The paring device has an inner diameter that
is smaller than the diameter of the core itself. The
aquifer material is extruded out of the core barrel
used for sampling and over the sterile paring device
to strip away the potentially contaminated material.
For anaerobic aquifers, this field paring procedure is
performed inside plastic anaerobic glovebags while
the latter is purged with nitrogen to minimize
exposure of the microflora to oxygen (Beeman and
Suflita, 1987). Samples received in this manner are
termed "aseptically obtained" and are suitable for
microbiological analysis.
Metabolic Activity in Aquifers
Many of the studies cited in Table 4 employed
aseptically obtained aquifer material and, thus,
provide direct and conclusive evidence for the
existence of subsurface bacteria. Several questions
immediately become apparent: (1) Are the indig-
enous ground-water organisms metabolically active;
(2) How diverse is their metabolism; (3) What factors
serve to limit and/or stimulate the growth and
metabolism of these organisms; and (4) Can one take
advantage of the inherent metabolic versatility of
aquifer communities to remediate contaminated
areas?
The answers to the first two questions can be gleaned
from a review of the scientific literature (Table 5).
Even though many of the studies cited in the table
were conducted prior to the development of aseptic
sampling procedures, the list does illustrate that
subsurface microbial activity can be detected, and
major and minor elements may potentially be
recycled in subsurface ecosystems. The metabolic
processes referred to in the table include a variety of
aerobic and anaerobic carbon transformations, many
of which are pertinent to the biodegradation of
aquifer contaminants. In this respect, the recent
reviews by Ghiorse and Wilson (1988) and Lee, et al.
(1988) are very illuminating. Other metabolic
processes noted are those required for the cycling of
nitrogen, sulfur, iron, and manganese.
The various metabolic processes listed in Table 5 are
not mutually exclusive. As shown schematically in
Figure 70, as labile organic matter enters an
oxygenated aquifer, microbial metabolism will likely
degrade the contaminating substrate. That is, the
indigenous microorganisms utilize the pollutant as
an electron donor to support heterotrophic microbial
respiration. The aquifer microbiota use oxygen as a
co-substrate and as an electron acceptor to support
their respiratory activities. This demand on oxygen
often results in its depletion and the establishment of
anaerobic conditions. When oxygen becomes limit-
ing, aerobic respiration slows, but other groups of
microorganisms may then come into play and
continue to degrade the contaminating organic
matter. Under conditions of anoxia, anaerobic
bacteria can use organic chemicals or several inor-
ganic anions as alternate electron acceptors.
Nitrate present in the ground water, as indicated in
Figure 70, generally is not rapidly depleted until
oxygen is utilized. Organic matter can still be
metabolized, but, instead of oxygen, nitrate serves as
the terminal electron acceptor during the process of
denitrification. Similarly, sulfate can serve as a
terminal electron acceptor when nitrate is limiting.
This information supports microbial metabolism
linked to sulfate-reduction. When this occurs, hydro-
gen sulfide often can be detected in the ground water
as a metabolic endproduct. When very highly
reducing conditions prevail in aquifers, carbon
dioxide also can serve as an electron acceptor and
metabolism is linked to methane formation. Some-
times a spatial separation of dominant metabolic
processes can occur in aquifers, depending on the
availability of electron acceptors, the presence of
suitable microorganisms, and the energy benefit of
70
-------
Table 5. Selected Microbial Metabolic Processes and Oxygen Requirements
Metabolic Process Oxygen Requirement
in Subsurface Environments
Reference
I. Biodegradation of Organic Pollutants
a. Petroleum
Hydrocarbons
b. Alkylpyridines
c. Creosote
Chemicals
d. Coal Gasification
Products
e. Sewage Effluent
f. Halogenated
Organic Compounds
g. Nitrilotriacetate (NTA)
h. Pesticides
II. Nitrification
III. Denitrification
IV. Sulfur Oxidation
V. Sulfur Reduction
VI. Iron Oxidation
VII. Iron Reduction
VIII. Manganese Oxidation
IX. Methanogenesis
Aerobic
Aerobic/Anaerobic
Aerobic/Anaerobic
Aerobic
'Aerobic
Aerobic/Anaerobic
Aerobic/Anaerobic
Aerobic/Anaerobic
Aerobic
Anaerobic
Aerobic
Anaerobic ,
Aerobic
Anaerobic
Aerobic
Anaerobic
Jamison, etal., 1975
Lee and Ward, 1985
McCarty, etal., 1984
Raymond, et al., 1976
Roberts, etal., 1980
Wilson, et al., 1983
Wilson, et al., 1985
Rogers, et al., 1985
Ehrlich, etal., 1983
Smolensk! and Suflita, 1987
Wilson, etal., 1985
Humenick, etal., 1982
Aulenbach, etal., 1975
Godsey and Ehrlich, 1978
Harvey, etal., 1984'
Gibson and Suflita, 1986
McCarty, etal., 1984
Suflita and Gibson, 1985
Suflita and Miller, 1985
Ward, 1985
Wilson, etal., 1983
Wood, etal., 1985
Ventullo and Larson, 1985
Ward, 1985
Gibson and Suflita, 1986
Suflita and Gibson, 1985
Ventullo and Larson, 1985
Ward, 1985
Barcelona and Naymic, 1984
Idelovitch and Michail, 1980
Preul, 1966
Ehrlich, etal., 1983
Lind, 1975
Ward, 1985
Olson, etal., 1981
Beeman and Suflita, 1987
Bastin, 1926
Hvid-Hansen, 1951
Jacks, 1977
Olson, etal., 1981
van Beek, etal., 1962
Olson, etal., 1981
Hallburg and Martinell, 1976
Godsey and Ehrlich, 1978
Ehrlich, et al., 1983
Hallburg and Martinell, 1976
Beeman and Suflita, 1987
Belyaev and Ivanov, 1983
Davis, 1967
Gibson and Suflita, 1986 ,,
Godsey and Ehrlich, 1978
Suflita and Miller, 1985
van Beek, et al., 1962
71
-------
GROUNDWATER FLOW
CO
»-
2
LU
ID
o
LU
+ 10
• 0
UJ
a
-10
CHEMICAL SPECIES
ELECTRON ACCEPTORS
ACETATE -CO,
SO'
CO,
BIOLOGICAL CONDITIONS
AEROBIC
HETEROTROPHIC
RESPIRATION
DENTTRFICATICN
SULFATE
RESPIRATION
METHANOGEhCSIS
Figure 70.
Mfcroblally mediated changes in chemical species, redox conditions, and spatial regions favoring different types
of metabolic processes along the flow path of a contaminant plume (adapted from Bouwer and McCarty, 1984):
the metabolic process to the catalyzing microbial
communities. As organic matter is transported in a
plume, a series of redox zones can be established
which range from highly oxidized to highly reduced
conditions. The biodegradation potential available
under these various conditions and the expected
rates of metabolism will be very different in each
instance.
The biodegradation of a known ground-water
contaminant, p-cresol, can be used as an example
(Figure 71). This compound is reported to be
72
-------
OH
COO"
Redox Conditions Biodegradability Lag Time Relative Rate Ref.
Aerobic
Denitrifying
Sulfate Reducing
(.Methanogenic
1
2
3,4
4,5,6
(1) Hopper, 1976, 1978^; (2) Bossert.& Young, 1986; (3) Bak &'Widdel, 1986;
(4) Smolenski & Suflita, 1-987; (5) Godsy et al., 1983; (6) Senior & Balba,
1984. :.;,,•
Figure 71. Proposed degradation pathway for p-cresol under various redox conditions.
degradable under aerobic, dinitrifying, sulfate-
reducing, and methanogenic conditions. Moreover,
the initial stages of the metabolism of p-cresol are
similar under the various redox conditions. Under
aerobic conditions, the initial conversion of the
substrate to aromatic alcohol requires molecular
oxygen. However, oxygen inhibits the anaerobic
metabolism of p-cresol, so, presumably, a water
molecule is employed to effect this transformation
during the catalysis of this substrate under the other
conditions. Comparing the length of the lag time
before the onset of biodegradation and the rate of
metabolism once it begins, aerobic decomposition is
relatively fast, especially compared to methanogenic
incubation conditions. Thus, if rapid biotrans-
formation is the goal, then aerobic incubation
conditions should be maintained in an aquifer in any
sort of a bioremediation strategy. However, consider-
ing the often slow rates of ground-water movement,
even anaerobic removal may effectively compete
with the faster aerobic metabolism of this substrate.
This does not imply that aerobic transformation
mechanisms are always faster'than the anaerobic
counterparts. Some compounds may biodegrade at
faster rates when anaerobic conditions prevail. For
example, several halogenated aliphatic substrates
prove amenable to microbial metabolism under
reducing conditions, but persist in comparable
aerobic incubations (Bouwer and McCarty, 1984;
Bouwer, et al., 1981). Similarly, the haloaromatic
pesticide 2,4,5-T often is considered recalcitrant
under aerobic conditions. Although specialized,
genetically manipulated microorganisms now exist
that aerobically degrade this substance (Kilbane, et
al., 1982), 2,4,5-T can be completely mineralized
without genetic manipulations when incubated
under methanogenic conditions (Gibson and Suflita,
1986).
When assessing microbial metabolism in aquifers,
the existing scientific literature must be used as a
guidepost for determining the types of biotrans-
formations to expect. For some substances, existing
information will be available and instructive; how-
ever, for more exotic substances, the available
literature will prove disappointingly small. In either
case, laboratory experimentation is required to eval-
uate a pollutant's susceptibility to biotransforma-
tion.
Microcosms
Even when the bulk of scientific literature indicates
that a pollutant chemical is likely to biodegrade
easily in most environments, an assay should be used
to confirm that contention. This is necessary
because, while a specific ground-water contaminant
73
-------
may be the same as that encountered in other
environments, the particular physical, chemical, and
hydrological site characteristics can be quite
different and can influence biodegradation processes.
A reliable tool or method is required for evaluating
the susceptibility of a ground-water contaminant to
biodegradation. In addition, the experimental sys-
tem should be scientifically rigorous, yet flexible
enough to facilitate a critical evaluation of various
bioremediation strategies. In this respect, the use of
microcosms seems to hold great promise. Some
examples of microcosm technology are discussed by
Pritchard (1981) and Pritchard and Bourquin (1984).
Also, Wilson and Noonan (1984) have analyzed the
specific application of microcosms to aquifer systems.
The term microcosm means different things to
different individuals. In the strictest sense, a
microcosm is:
...a calibrated laboratory simulation of a portion
of a natural environment in which environ-
mental components, in as undisturbed a condi-
tion as possible, are enclosed within definable
physical and chemical boundaries and studied
under a standard set of laboratory conditions
(Pritchard, 1981).
Implicit in this definition is that microcosms may be
used as surrogates for ground-water field studies
without the associated logistical, financial,
administrative, and regulatory problems. In addi-
tion, when properly employed, microcosms incor-
porate the site-specific variables of the area under
investigation and allow for the assessment of risk
and prediction of the transport and fate charac-
teristics of a contaminant. Further, such systems
help to evaluate the waste assimilative capacity of
an environment (Pritchard and Bourquin, 1984).
This evaluation can be achieved without a detailed
knowledge of those endogenous environmental
factors that may stimulate or retard microbial
metabolism.
The design of microcosms depends a great deal on the
nature of the question being posed (Wilson and
Noonan, 1984; and Pritchard and Bourquin, 1984).
Microcosms range from simple batch incubation
systems to large and complex flow-through devices,
and all can be used in a variety of studies on the
transport and fate of subterranean contaminants
influenced by microorganisms. Microcosms are used
to help identify biodegradable pollutants and those
pollutants that tend to persist. Often, by following
the fate of a pollutant in a microcosm, the predom-
inant pathways of biotic or abiotic transformation
can be described. Lastly, the decay of a particular
contaminant in a microcosm may relate to the rate of
biotransformation in situ without relying on indirect
measures of microbial biodegradation for rate predic-
tions such as the enumeration of catalytic micro-
organisms.
This last point can be illustrated with an example.
An assay for the anaerobic biodegradation of the
pesticide 2,4,5-T was conducted with microcosms
made of sediments and ground water sampled from
two distinct sites within the same aquifer (Gibson
and Suflita, 1986). Microorganisms from one site
were able to completely mineralize this substrate
while parallel experiments revealed that the
pesticide persisted in microcosms made from the
other site. Further experimentation showed that the
requisite microorganisms were present at both sites,
but their activity at the second site was at least
partially inhibited by locally high levels of sulfate in
the ground water. Predictions of the rate of substrate
utilization in the sulfate-rich ground water would
have been overly optimistic if based primarily on
assays for the number of degrading microorganisms.
Conversely, predictions of transport and fate
behavior based on microcosm experimentation would
likely prove more realistic, even without the
appreciation for the role of sulfate in a portion of the
aquifer. This example also illustrates how site-
specific aquifer characteristics can drastically influ-
ence the evaluation of biodegradation results.
Microcosms possess several experimental advan-
tages. For example, microcosms are replicable and
allow appropriate controls to be employed. This
advantage cannot be overstated since it allows the
resulting data to be statistically evaluated. Further,
such experiments can meet the usual requirements
for scientific rigor. While environmental simulation
is a goal of microcosms, such systems may be subject
to controlled perturbations of their chemical or
physical parameters. In that manner, the influence
of such perturbations can be quantitatively evalu-
ated for effects on biotransformation processes.
Similarly, the trophic structure of microcosms can be
varied in order to study complex biological inter-
actions. However, this advantage may not exist since
microeucaryotes tend to be of minor importance in
most aquifers examined thus far. Microcosms are
accessible and containable laboratory tools with
which the experimenter controls inputs (water,
substrate, nutrients, etc.) as well as exports (cells,
metabolites, etc.). When properly employed, micro-
cosms avoid field pollution, yet provide a time-
efficient way of evaluating the likely fate of environ-
mental contaminants.
Conversely, microcosms also suffer from several
limitations that should not be ignored at the onset of
experimentation and during an evaluation of the
resulting data. One limitation is that the start-up
and operating costs for complex microcosms can be
74
-------
high. Also, microcosms can disturb the normal struc-
tural and functional features of an ecosystem and
generally possess abnormally high surface to volume
ratios. The simple sampling and containerization of
microbial communities within altered physical
limits causes the initiation of rapid population shifts
that may influence biodegradation estimates. How-
ever, the advantages of microcosm techniques out-
weigh the inherent limitations (Pritchard and
Bourquin, 1984), and such studies can provide useful
and environmentally reliable information.
Extrapolation of Metabolic Information
Although microcosms provide a tool that can be used
to address some of the questions regarding metabolic
proficiency of ground-water microorganisms, other
questions remain. For example, is the information
collected from microcosm studies reliable, and can
the metabolic information obtained from microcosm
studies be reliably extrapolated to terrestrial
subsurface environments? These questions stem
from a consideration of the basic principles that
generally govern biodegradation. Although subsur-
face microbes may or may not prove to be unusual,
metabolism tends to be a unifying feature of life. As
Dagley (1984) has pointed out, diverse life forms
often exhibit similar metabolic pathways, and
information pertinent to even human physiology
often can be obtained from microbiological studies
(Dagley, 1984). Therefore, metabolic principles
gleaned from the study of xenobiotic compound
metabolism in surface habitats are useful for
subsurface site evaluations and the design of aquifer
remediation strategies. The existing literature
provides an excellent scientific foundation that can
be used to make cautious extrapolations of metabolic
principles observed in surface environments to those
that regulate microbial activities in the subsurface.
Microorganisms play prominent roles in the
transformation and degradation of contaminating
organic chemicals in virtually every major habitat
except the atmosphere. Microbial communities in
nature are able to metabolize many types and
quantities of synthetic organic compounds (Alex-
ander, 1981; Kobayashi and Rittman, 1982). At the
extremes, some xenobiotic compounds can supply the
carbon and energy requirements associated with
microbial growth, while other substrates tend to
resist significant biotransformation. Still other
chemicals are "cometabolized" to form degradation
products that are usually structurally similar to the
parent molecule (Alexander, 1981). By definition,
the latter process does not result in growth of the
catalyzing microorganism, but does result in the
transformation of a substrate such that the molecule
can be available for subsequent utilization by other
microorganisms. In an ecological sense, it is common
that several microbes act in a combination to effect
some overall degradation of a substrate (Slater and
Lovatt, 1984). In anaerobic ecosystems, this con-
certed metabolic activity is essential for the normal
cycling of carbon and energy on earth (Mclnerney
and Bryant, 1981).
Knowledge of the mechanisms of energy and
materials cycling in the terrestrial subsurface,
however, is limited and reflects the difficulties in
sampling and studying a complex and unfamiliar
environment (McNabb and Mallard, 1984). As
pointed out earlier, the likely metabolic fate of
contaminating chemicals sometimes can be predicted
if the governing microbial ecology and the limits of
extrapolation are understood.
Recalcitrance
Researchers have learned that there are limits to the
metabolic versatility of microorganisms. Many xeno-
biotic substrates are transformed so slowly under
most conditions that some degradation of environ-
mental quality occurs. These chemicals are referred
to as recalcitrant. However, recalcitrance or resis-
tance to biodegradation is not a feature strictly
associated with anthropogenic chemicals'. Many
naturally occurring materials also can persist for
long periods of time (Alexander, 1965 and 1973). For
example, entire portions of human corpses (i.e., hair)
were recovered from various environments (peat
bogs, desert cemeteries, etc.) and estimated to be
hundreds to thousands of years old. Similar ages
have been ascribed to proteolytic enzymes recovered
from permafrost soils. Microbial spores can persist
for tens of thousands of years. These spores are
useful to help "date" geological deposits. Finally,
most of the world economies are based on petroleum
deposits that have managed to persist for many
millennia within subterranean environments.
In aquifers, as in other environments, organic matter
decomposition should be considered in the context of:
(1) existing environmental conditions; (2) physiology
of the requisite microorganisms; and (3) chemical
structure of the particular substance(s) under consid-
eration. The general principles of these interrelated
topics are considered below; detailed reviews can be
found in Alexander's publications (1965 and 1973).
Environmental Barriers to
Biodegradation
In order to grow, microorganisms need a suitable
physical and chemical environment. Extremes of
temperature, pH, salinity, osmotic or hydrostatic
pressures, radiation, free water limitations, contami-
75
-------
nant concentration, and/or the presence of a heavy
metal or other toxicant materials can adversely
influence and even limit the rate of microbial growth
and/or substrate utilization. Often, two or more
environmental factors interact to limit microbial
decomposition processes, and, in fact, environmental
barriers can act to render normally labile substances
persistent. This explanation was used by Woods Hole
researchers (Jannasch, et al., 1971) who found food
substances from an accidentally sunken and
subsequently recovered submarine to be almost
preserved after 10 months exposure to deep-sea
conditions. However, when the food was incubated in
sterile seawater at in situ temperatures (3°C), the
materials putrefied after only a few weeks. The
authors suggested that the hydrostatic pressures of
the sea (150 ATM) effectively raised the minimal
temperature necessary for microbial proliferation.
Once this increase exceeded incubation
temperatures, microbial activity slowed 10 to 100
fold.
Recognition of the nature of the limiting environ-
mental factor(s) and a consideration of its practical
application in subsurface environments will help
dictate which type of bioreclamation procedures to
use. For example, the temperature of aquifers prob-
ably could not be significantly altered to stimulate in
situ microbial growth and metabolism. However, the
same is not true for a surface biological treatment
process.
Physiological Barriers to Biodegradation
In addition to the immediate environment, various
microbial physiological factors can influence the
biotransformation of pollutant chemicals. Like all
forms of life, the requisite microorganisms present in
an aquifer are primarily composed of carbon,
hydrogen, oxygen, nitrogen, phosphorus, sulfur, and
a variety of trace elements. These substances are
required in varying degrees for aquifer micro-
organisms to proliferate. Aquifer microorganisms
can utilize such substances to the point where one or
more of the requirements are exhausted and
effectively limit further microbial growth or
metabolic activity. For aquifer remediation efforts to
succeed, these materials must already be present in
the aquifer or be supplied in the proper form.
Ideally, the organic pollutants in the aquifer repre-
sent an appropriate supply of carbon and energy
necessary for heterotrophic microbial growth.
However, that supply can be too high or too low. Too
high a substrate concentration can limit microbial
metabolism due to the toxicity of the substrate to the
requisite microflora. In contrast, ground-water con-
taminants also can be present at concentrations that
are too low to allow microbial response and/or may
not be suitable growth substrates. Growth and
energy sources need not come from the same carbon
substrate. The growth and metabolism of degrading
microbes sometimes can be stimulated by providing
them with a non-harmful primary carbon substrate
so that the rate and extent of pollutant decompo-
sition can be proportionally increased (McCarty,
1985; McCarty, et al., 1981; McCarty, et al., 1984).
A chemical also will be metabolized poorly if it is
unable to enter microbial cells. This may occur with
either natural or anthropogenic polymers. While the
monomeric units may be inherently amenable to
microbial destruction, the larger molecular weight
polymers persist because they often fail to gain
access to intracellular catabolic enzymes. A sub-
strate also will persist if it fails to derepress the
enzymes necessary for its destruction. It may be
possible to induce the appropriate enzymes with an
alternate chemical compound. Occasionally, initial
biochemical reactions result in metabolites that tend
to inhibit the degradation of even the parent
molecule and can adversely affect the biodegradation
of some pollutants.
Lastly, the absence of other necessary micro-
organisms can limit the destruction of a contami-
nant. Often, several microbial groups are needed for
the destruction of a pollutant. In anaerobic environ-
ments, this type of relationship is a prerequisite for
completion of the carbon cycle (Mclnerney and
Bryant, 1981). The anaerobic mineralization of
organic matter is critically dependent on obligate
microbial consortia; if any of the individual members
of a consortium are absent, the biodegradation of the
parent material effectively ceases.
Chemical Barriers to Biodegradation
One of the most important factors that influence the
degradation of a contaminant in aquifers is the
contaminant's structure, which can dictate the
pollutant's physical state (i.e., soluble, adsorbed,
conjugated, etc.) and thereby alter its tendency to
biodegrade. However, it is important to be specific
when referring to biodegradation. When a compound
undergoes primary attack, initial metabolic events
often result in reaction products with their own
environmental impact and persistence character-
istics. An example of this is the fate of the pesticide
carbaryl (Sevin). Carbaryl is widely used as a
substitute for DDT and is touted as being readily
degradable. However, this pesticide is known to have
a myriad of environmental fates (Figure 72). An
examination of Figure 72 reveals that somte
metabolic routes do lead to the ultimate conversion
of carbaryl to carbon dioxide, while others result in
the production of complex aromatic metabolites, the
fate of which is still unknown. Therefore, when
76
-------
OCONHCH3
OH
Figure 72. Proposed pathway of carbaryl degradation in soils and microbial cultures (adapted from Rajagopal, et al., 1984).
evaluating the scientific literature, the distinction
must be made between "biodegradation" and
"mineralization." The mere loss of a chemical from
an environment may or may not be a desirable
consequence of biotransformation processes. If bio-
degradation results in the production of undesirable
metabolites, it may be best to choose a non-biological
strategy for aquifer cleanup. The example above
illustrates the need to understand the overall fate of
a contaminating chemical in an aquifer and how
informed decision-making must be based on solid
metabolic information.
Aquifer pollutants may contain various chemical
linkages that tend to favor or hinder microbial
attack. However, broad generalizations on the biode-
gradability of various linkages tend to be of limited
value since substitute effects drastically alter the
77
-------
the molecule proved available to only a few organ-
isms and permitted minimal growth at best. Very
often, when a quaternary carbon atom occurs at the
terminal end of an alkane chain, the resulting
molecule is quite resistant to aerobic microbial
attack.
Structural effects also can be observed with
anaerobic microorganisms. A comparison of the rate
of aryl-reductive dehalogenation by an anaerobic
consortium of bacteria revealed that the type and
position of the halogen substituent influenced the
degradation of the resulting molecules. Table 7
shows that all brominated and iodinated benzoates
were transformed by these bacteria, but no fluori-
nated benzoates and only one chlorinated benzoate
susceptibility of even simple organic molecules to
biotransformation. The number, type, and position of
substituents must be considered when evaluating
the metabolic fate of particular contaminants in
aquifers.
The effect of branching is illustrated in Table 6
which is adapted from the work of McKenna and
Kallio (1964). Approximately 15 bacterial strains
from several different genera were evaluated for
their ability to use a series of structurally related
substrates. Cultures exhibiting profuse growth,
slight growth, or no growth on a particular substrate
were assigned a qualitative index of 2, 1, or 0,
respectively. All cultures grew well with 1-phenyl-
decane. The organisms most likely initiated their
attack on this molecule through the oxidation of the
terminal methyl group in order to convert the
molecule to a fatty acid derivative. Continued metab-
olism of the molecule would likely be by 0-oxidation
of the side chain. However, a single methyl substitu-
ent on the side chain drastically influenced the
susceptibility of the resulting structure to
metabolism by the test organisms. All the myco-
bacteria and nocardia were able to grow on 1-phenyl-
4-methyldecane. However, the pseudomonads and
micrococci were no longer able to do so. When the
substrate contained an "internal" quaternary carbon
atom, as is the case for l-phenyl-4,4-dimethyldecane,
Table 6. Microbial Growth Response to Phenyldecanes: 0 = No Growth; 1 = Questionable or Slight Growth; 2 = Moderate
or Abundant Growth (Adapted from McKenna and Kallio, 1964)
Table 7.
The Rate of Anerobic Monohalobenzoate
Metabolism Exhibited by an Enrichment of
Dehalogenating Bacteria (DeWeerd, et al., 1986)
Dehalogenation Rate (nmoles/L/hr)
Position
Cl
Bf
Ortho
Meta
Para
n.d.
0
0
0
4.63
0
1.20
3.70
0.05
0.50
0.89
0.66
Organism
M/crococcus cerificans
Pseudomonas aeruginosa
Mycobactenum phlei
M, fortuitum
M, rhodochrous
M. smegmatis
Nocardia opaca
N, rubra
N, erythropolis
N, polychromogenes
N, corallma
Strain
H.0.1-N
H.0.3
H.0.4
S-18.2
S-14.1
119JWF
191 JWF
Sol 20 JS
No. 451
No. 389
No. 382
No. 422
1 -Phenyldecane
2
2
2
2
2
2
2
2
2
?.
2
2
2
2
2
2
2
1 -Phenyl-4-methyldecane
' 0
0
• • o
,0
" °
0
0
0
2
2
2
: 2
2
2
2
2
2
i -rnenyi-q-,4-
dimethyldecane
0
0
0
0
0
0
0
0
0
0
0
1-
1
1 .
1
1
1
78
-------
were transformed. For transformed compounds,
slower rates of dehalogenation were correlated with
increasing halogen size. However, the rate of tirtKo,
meta, and para dehalogenation differed less with
increasing halogen size. Most importantly, the meta-
substituted halobenzoates always exhibited the
fastest rates of biodegradation when compared to
their isomeric counterparts. Thus, when halobenzo-
ates are considered, the meto-substituted molecules
are the most susceptible to anaerobic biotf ansforma-
tion. — :J%
These few examples illustrate that the chemical
structure of a contaminant also influences its sus-
ceptibility to biodegradation. Despite the difficulty
in generalization, the scientific literature shows that
the closer a contaminant structurally resembles a
naturally occurring compound, the better the
possibility that the former will be able to enter a
microbial cell, derepress the synthesis of catabolic
enzymes, and be converted by those enzymes to
central metabolic intermediates. In addition, biode-
gradation is less likely, but not entirely precluded,
for those molecules having unusual structural
features only infrequently encountered in nature.
Ease of biodegradation can be viewed as a continuum
ranging from very labile compounds to those that are
recalcitrant (Figure 73).
Labile
Recalcitrant
Structural
Analogs of
Natural Materials
Chemicals with
No Natural
Counterpart
Figure 73. The continuum of biodegradation ease.
When viewed in this context, it should not be
surprising that xenobiotic compounds tend to persist
in nature because microorganisms have not evolved
the necessary metabolic machinery to attack those
compounds. However, as a group, microorganisms
are nutritionally versatile, have the potential to
grow rapidly, and possess only a single copy of DNA.
Consequently, any genetic mutation or recombina-
tion is immediately expressed. If the alteration is of
adaptive significance, new species of microorganisms
can arise and proliferate. The polluted environment
supplies selection pressure for the evolution'of
organisms with novel metabolic potential. Ulti-
mately, the organisms may not only survive in the
polluted environment, but also may be capable of
growing at the expense of the contaminating
substance. .......
Bioremediation of Aquifers
Once an aquifer contaminant is recognized as being
susceptible to biodegradation, the goal of bioremedi-
ation/bioreclamation efforts is to utilize the meta-
bolic capabilities of the indigenous microflora to
destroy and eliminate that contaminant. Currently,
enhanced bioreclamation is the use of common
aquifer bacteria to degrade organic contaminants.
This practice generally does not require the innocu-
latidn of the terrestrial subsurface with foreign
bacteria.
Bioremediation attempts to impose particular
conditions in an aquifer that encourage microbial
proliferation and the development of desirable micro-
organisms. With knowledge of the chemical and
physical needs of microorganisms and the pre-
dominant metabolic pathways, a strategy can be
developed to stimulate biodegradation. Most often,
microbial activity is stimulated by supplying the
nutrients necessary for microbial growth. These
efforts can take place either above ground or in situ.
An example of above-ground bioreclamation is
schematically shown in Figure 74. In the illus-
tration, ground water and the associated contam-
inant are pumped to the surface through a series of
recovery wells. On the surface, the water is treated
in a series of steps, which include the supply of
appropriate nutrients. The treated (i.e., decontam-
inated) water is then reinjected into the aquifer.
Injection and recovery wells are positioned so that
they intersect a zone of contamination within the
aquifer along natural ground-water flow paths. If
excess nutrients are in the treated water, it is
conceivable that this system also will stimulate the
in situ biotransformation of the contaminants.
Moving the contaminant to the surface for treatment
is often impossible or impractical since even mildly
adsorbing chemical species may require many
decades of pumping before being reduced to suf-
ficiently low levels. In this case, an in situ approach
may be more feasible. Figure 75 shows two such
systems.
Figure 75A illustrates a system in which the
microbial nutrients are mixed with ground water
and circulated through the contaminated portion of
the aquifer via a series of injection and recovery
wells. In this case, air compressors are used to
impose aerobic conditions on the indigenous micro-
flora. Nutrients also can be circulated via an infiltra-
tion gallery as depicted in Figure 75B. This method
provides a potential mechanism for the microorgan-
isms to attack contaminants trapped in the pore
spaces of the unsaturated zone. Each of the examples
79
-------
DIRECTION OF GROUND WATER FLOW
outside water source
A
INJECTION SYSTEM
•* —
t
' treated water / /
A"\ y ^
r clean \ -, ^ ^
. pool ) S^\
/ t
M a »
^
• M M • a
V
ZONE OF CONTAMINATION
V I
^
contaminated
water T T
I I J 1
f w w ir
T T T T T T T t
r —
c
pump
1
1
1
1
1
1
^ RECOVERY
/
A
1
SYSTEM
4
1
1
J
J
i
1
Figure 74.
•
treatment system design for the biorestoration of a contaminated aquifer (adapted from Lee, et
discussed above presume an adequate understanding
and control of the hydrogeology of the area under
treatment. This topic will be dealt with more
specifically in Chapter 8.
Biorestoration approaches have several advantages
over other potential cleanup strategies. First, these
approaches can be used to treat some common
aquifer pollutants, as described in Chapter 8. Bio-
remediation is environmentally sound in that it
results in the complete destruction of a contaminant.
This is not to imply that the technology can remove
99.999 percent of the material, but rather, it results
m the complete change in the chemical properties of
the parent contaminant. Aquifer biorestoration
approaches utilize the indigenous microflora so that
mass inoculation efforts are often unnecessary.
There is insufficient scientific evidence to suggest
that the intentional addition of desirable micro-
organisms actually aided an aquifer biorestoration
effort (Lee, et al., 1988; Chapter 8). Finally,
compared to other technologies, biorestoration also is
economical.
However, biorestoration also suffers from several
drawbacks (Lee, et al., 1988). At some sites, many
types of wastes (acids, bases, and organic and
inorganic materials) are co-disposed and any of these
wastes alone or in combination could inhibit
microbial growth and metabolism. Incomplete
degradation of some substances may lead to taste
and odor problems in the ground water. Also, the
operation of biorestoration programs can be
maintenance- and analytically intensive. Further, to
date, the technology has generally been limited' to
aquifers possessing high permeabilities. Despite
these limitations, the cost of competing technologies
will likely insure that bioreclamation efforts will
continue and assume an increasingly larger role as a
viable aquifer cleanup strategy.
References
Alexander, M., 1965. "Biodegradation: Problems of
Molecular Recalcitrance and Microbial Fallibility."
Adv. Appl. Microbiol., Vol. 7, pp. 35-80.
Alexander, M., 1973. "Nonbiodegradable and Other
Recalcitrant Molecules." Biotech. Bioeng Vol 15
pp. 611-647. ' '
Alexander, M., 1981. "Biodegradation of Chemicals
of Environmental Concern." Science, Vol 211 nn
132-138. • ' PP'
Aulenbach, D. B., N. L. Clesceri, and T. J. Tofflemire,
1975. "Water Renovation by Discharge into Deep
80
-------
To Sewer or
Recirculate
Air
Compressor
Nutrient
Addition
Tank
Production
Well
Coarse Sand
Water Table-
^_ — Spilled Materials __
Water Supply
Injection
Well
— Sparger
•. • . Clay .' • '. • '
Air Compressor or
Hydrogen Peroxide Tank
Nutrient Addition
-Infiltration Gallery
^Trapped Hydrocarbons
Recirculatec
Water 8 Nutrients/
Monitoring Well
Water/"-.^
Table
Recovery Well
Figure 75. Two designs for the in situ remediation of contaminated aquifers (adapted from Lee, et al., 1988).
Natural Sand Filters." Proceedings of AIChE
Conference, May 4-8,1975, Chicago, IL.
Bak, F. and F. Widdel, 1986. "Anaerobic Degrada-
tion of Phenol and Phenol Derivatives by Desulfo-
bacterium Phenolicum sp. nov." Arch. Microbiol.,
Vol. 146, pp. 177-180.
Balkwill, D. L. and W. C. Ghiorse, 1985. "Character-
ization of Subsurface Bacteria Associated with Two
Shallow Aquifers in Oklahoma." Appl. Environ.
Microbiol., Vol. 50, pp. 560-588.
Balkwill, D. L., T. E. Ruzinski, and L. E. Casida,
1977. "Release of Microorganisms from Soil with
Respect to Transmission Electron Microscopy
Viewing and Plate Counts." Antonie Van Leeu-
wenhoekJ. Microbiol. Serol., Vol. 43, pp. 73-87.
Barcelona, M. J. and T. G. Naymik, 1984. "Dynamics
of a Fertilizer Plume in Ground Water." Environ.
Sci. Technol., Vol. 18, pp. 257-261.
Bastin, E. S., 1926. "The Problem of the Natural
Reduction of Sulfates." Bull. Amer. Assoc. Petrol.
Geol., Vol. 10, pp. 1270-1299.
81
-------
Beeman, R. E. and J. M. Suflita, 1987. "Microbial
Ecology of a Shallow Unconfined Ground-water
Aquifer Polluted by Municipal Landfill Leachate."
Microb. Ecol., Vol. 14, pp. 39-54.
Belyaev, S. S. and M. V. Ivanov, 1983. "Bacterial
Methanogenesis in Underground Waters." Ecol.
Bull, Vol. 35, pp. 273-280.
Bitton, G. and C. P. Gerba, 1984. "Ground-water
Pollution Microbiology: The Emerging Issue." In:
Groundwater Pollution Microbiology, Bitton, G.,
and C. P. Gerba, Editors. John Wiley & Sons, New
York, pp. 1-7.
Bossert, I. D. and L. Y. Young, 1986. "Anaerobic
Oxidation of p-cresol by a Denitrifying Bacterium."
Appl. Environ. Microbiol, Vol. 52, pp. 1117-1122.
Bouwer, E. J. and P. L. McCarty, 1984. "Modeling of
Trace Organics Biotransformation in the Sub-
surface." Groundwater, Vol. 22, pp. 433-440.
Bouwer, E. J., B. E. Ritman, and P. L. McCarty,
1981. "Anaerobic Degradation of Halogenated 1-
and 2-Carbon Organic Compounds." Environ. Sci.
Tecknol, Vol. 15, pp. 596-599.
Craun, G. P., 1984. "Health Aspects of Ground-water
Pollution." In: Ground-water Pollution Microbi-
ology, Bitton, G., and C. P. Gerba, Editors. John
Wiley & Sons, New York, pp. 135-195.
Council on Environmental Quality, 1981. "Contam-
ination of Ground Water by Toxic Organic
Chemicals." U.S. Government Printing Office,
Washington, DC.
Dagley, S., 1984. "Introduction." In: Microbial
Degradation of Organic Compounds, Gibson, D. T.,
Editor. Marcel Dekker, Inc., New York, pp. 1-10.
Davis, J.B., 1967. Petroleum Microbiology. Elsevier
Publishing Co., Amsterdam, p. 604.
DeWeerd, K., et al., 1986. "The Relationship
Between Reductive Dehalogenation and Other
Aryl Substituent Removal Reactions Catalyzed by
Anaerobes." FEMS Microb. Ecol., Vol. 38, pp. 331-
339.
Ehrlich, G. G., et al., 1983. "Microbial Ecology of a
Cresote-contaminated Aquifer at St. Louis Park,
Minnesota." Dev. Ind. Microbiol., Vol. 24, pp. 235-
245.
Ghiorse, W. C. and D. L. Balkwill, 1983. "Enumer-
ation and Morphological Characterization of
Bacteria Indigenous to Subsurface Environments."
Dev. Ind. Microbiol., Vol. 24, pp. 213-224.
Ghiorse, W. C. and D. L. Balkwill, 1985.
"Microbiological Characterization of Subsurface
Environments." In: Ground-water Quality, C. H.
Ward, W. Giger and P. L. McCarty, Editors. John
Wiley & Sons, Inc., New York, pp. 536-556.
Ghiorse, W. C. and J. T. Wilson, 1988. "Microbial
Ecology of the Terrestrial Subsurface." Adv. Appl.
Microbiol., Vol. 33, pp. 107-172.
Gibson, S. A. and J. M. Suflita, 1986. "Extrapolation
of Biodegradation Results to Groundwater
Aquifers: Reductive Dehalogenation of Aromatic
Compounds." Appl. Environ. Microbiol., Vol. 52,
pp. 681-688.
Godsey, E. M. and G. G. Ehrlich, 1978.
"Reconnaissance for Microbial Activity in the
Magothy Aquifer, Bay Park, New York, Four
Years After Artificial Recharge." J. Res. U.S. Geol.
Survey, Vol. 6, pp. 829-836.
Godsey, E. M., D. F. Goerlitz, and G. G. Ehrlich,
1983. "Methanogenesis of Phenolic Compounds by
a Bacterial Consortium from a Contaminated
Aquifer in St. Louis Park, Minnesota." Bull.
Environ. Contam. Toxicol., Vol. 30, pp. 261-268.
Hallburg, R. O. and R. Martinell, 1976. "Vyredox - In
Situ Purification of Ground-water." Ground-water,
Vol. 14, pp. 88-93.
Harvey, R. W., D. L. Smith, and L. George, 1984.
"Effect of Organic Contamination Upon Microbial
Distribution and Heterotrophic Uptake in a Cape
Cod, Massachusetts Aquifer." Appl. Environ.
Microbiol., Vol. 48, pp. 1197-1202.
Hopper, D. J., 1976. "The Hydroxylation of p-cresol
and Its Conversion to p-hydroxybenzaldehyde in
Pseudomonas putida." Biochem. Biophys. Res.
Commun., Vol. 69, pp. 462-468.
Hopper, D. J., 1978. "Incorporation of [18Q] Water in
the Formation of p-hydroxybenzyl Alcohol by the
p-cresol Methylhydroxylase from Pseudomonas
putida." J. Biochem., Vol. 175, pp. 345-347.
Humenick, M. J., L. N. Bitton, and C. F. Mattox,
1982. "Natural Restoration of Ground Water in
UCG."InSitu, Vol. 6, pp. 107-125.
Hvid-Hansen, N., 1951. "Sulphate-reducing and
Hydrocarbon-producing Bacteria in Ground
Water." Act. Pathol. Microbiol. Scand., Vol. 29, pp.
314-334.
Idelovitch, E. and M. Michail, 1980. "Treatment
Effects and Pollution Dangers of Secondary
Effluent Percolation to Groundwater." Prog. Wat.
Tech., Vol. 12, pp. 949-966.
Jacks, G., 1977. "The 'Amber River.' An Example of
Sulfate Reduction." Proc. Second Int. Symp. Water-
Rock Interaction. Strasburg, Aug. 17-25, Vol. I, pp.
259-266.
Jamison, V. W., R. L. Raymond, and J. O. Hudson,
1975. "Biodegradation of High-octane Gasoline in
Groundwater." Dev. Ind. Microbiol., Vol. 16, pp.
305-312.
82
-------
Jannasch, H., et al., 1971. "Microbial Degradation of
Organic Matter in the Deep Sea." Science. Vol. 171,
pp. 672-675.
Keswick B .H.} 1984. "Sources of Ground-water
Pollution." In: Groundwater Pollution Micro-
biology, Bitton, G., and C. P. Gerba, Editors. John
Wiley & Sons, New York, pp. 39-64.
Kilbane, J. J., et al., 1982. "Biodegradation of 2,4,5-
trichlorophenoxyacetic Acid by a Pure Culture of
Pseudomonas cepacia." Appl. Environ. Microbiol.,
Vol. 44, pp. 72-78.
Kobayashi, H. and B. E. Rittmann, 1982. "Microbial
Removal of Hazardous Organic Compounds."
Environ. Sci. TechnoL, Vol. 16, pp. 170a-183a.
Kuznetsov, S. I., N. V. Ivanov, and N. N. Lyalikova,
1963. "The Distribution of Bacteria in Ground-
waters and Sedimentary Rocks." In: Introduction to
Geological Microbiology, C. Oppenheimer, Editor.
McGraw-Hill Book Co., New York.
Ladd, T. I., et al., 1982. "Heterotrophic Activity and
Biodegradation of Labile and Refractory Com-
pounds by Ground Water and Stream Microbial
Populations." Appl. Environ. Microbiol., Vol. 44,
pp. 321-329.
Lee, M.'D., et al., 1988. "Biorestoration of Aquifers
Contaminated with Organic Compounds." CRC
Crit. Rev. Environ. Control, Vol. 18, pp. 29-89.
Lee, M. D. and C. H. Ward, 1985. "Biological
Methods for the Restoration of Contaminated
Aquifers." Env. Toxicol. Chem., Vol. 4, pp. 721-726.
Lind, A. M., 1975. "Nitrate Reduction in the
Subsoil." Proc. Int. Assoc. Water Pollut. Res.
Copenhagen, Aug. 18-20, Vol. 1,-p. 14.
MacKay, D. M., P. V. Roberts, and I. A. Cherry, 1985.
"Transport of Organic Contaminants in Ground
Water." Environ. Sci. TechnoL, Vol. 19, pp. 384-392.
McCarty, P. L., 1985. "Application of Biological
Transformations in Groundwater." Proc. Second
Int. Conf. Ground-water Quality Res., Durham, N.
N. and A. E. Redelfs, Editors. Natl. Cntr. Ground-
water Res., Stillwater, OK, pp. 6-11.
McCarty, P. L., M. Reinhard, and B. E. Rittmann,
1981. "Trace Organics in Groundwater." Environ.
Sci. TechnoL, Vol. 15, pp. 47-51.
McCarty P. L., B. E. Rittmann, and E. J. Bouwer,
1984. "Microbiological Processes Affecting Chemi-
cal Transformations in Groundwater." In: Ground-
water Pollution Microbiology, Bitton, G., and C. P.
Gerba, Editors. John Wiley & Sons, New York.
Mclnerney, M. J. and M. P. Bryant, 1981. "Basic
Principles of Bioconversions in Anaerobic Diges-
tion and Methanogenesis." Biomass Conversion
Processes for Energy and Fuels, Soferr, S. S. and O.
R. Zaborsky, Editors. Plenum Publ. Corp., New
York, pp. 277-296.
McKenna, E. J. and R. E. Kallio, 1964. "Hydrocarbon
Structure: Its Effect on Bacterial Utilization of
Alkanes." In: Principles and Applications in
Aquatic Microbiology, H. Heukelekian and N. C.
Dondero, Editors. John Wiley & Sons, New York,
pp. 1-14.
McNabb, J. F. and W. J. Dunlap, 1975. "Subsurface
Biological Activity in Relation to Groundwater
Pollution." Groundwater, Vol. 13, pp. 33-34.
McNabb, J. F. and G. E. Mallard, 1984. "Micro-
biological Sampling in the Assessment of
Groundwater Pollution." Groundwater Pollution
Microbiology, Bitton, G. and C. P. Gerba, Editors.
John Wiley & Sons, New York, pp. 235-260.
Olson, G. S., et al., 1981. "Sulfate-reducing Bacteria
from Deep Aquifers in Montana." Geomicrobiol J.,
Vol. 2, pp. 327-340.
Olson, G. J., G. A. McFeters, and K. L. Temple, 1981.
"Occurrence and Activity of Iron and Sulfur-
oxidizing Microorganisms in Alkaline Coal Strip
Mine Spoils." Microb. EcoL, Vol. 7, pp. 40-50.
Page, G. W., 1981. "Comparison of Groundwater and
Surface Water for Patterns and Levels of
Contamination by Toxic Substances." Environ.
Sci. TechnoL, Vol. 15, pp. 1475-1481.
Preul, H. C., 1966. "Underground Movement of
Nitrogen." Adv. Water Pollut. Res.: Proc. Third
Int. Conf., pp. 309-328.
Pritchard, P. H., 1981. "Model Ecosystems." In:
Environmental Risk Analysis for Chemicals, R. A.
Conway, Editor. Van Nostrand Reinhold, New
York, pp. 257- 353.
Pritchard, P. H. and A. W. Bourquin, 1984. "The Use
of Microcosms for Evaluation of Interactions
Between Pollutants and Microorganisms." In: Adv.
Microbial. Ecology, K. C. Marchall, Editor.
Plenum Publishing Corporation, Vol. 7, pp. 133-
215.
Pye, V. I. and R. Patrick, 1983. "Groundwater Con-
tamination in the United States." Science, Vol.
221, p. 713.
Rajagopal, B. S., et al., 1984. "Effect and Persistence
of Selected Carbamate Pesticides in Soil." Res.
Rev., Vol. 93, pp. 1-199.
Raymond, R. L., V. W. Jamison, and J. O. Hudson,
1976. "Beneficial Stimulation of Bacterial Activity
in Ground Water Containing Petroleum Products."
Aice Symposium Series, Vol. 73, pp. 390-404.
Roberts, P. V., et al., 1980. "Organic Contaminant
Behavior During Groundwater Recharge." J. Water
Pollut. Control Fed., Vol. 52, pp. 161-172.
83
-------
Rogers, J. E., et al., 1985. "Microbial Transformation
of Alkylpyridines in Ground-water." Wat. Air Soil
Poll., Vol. 24, pp. 443-454.
Senior, E. and M. T. M. Balba, 1984. "The Use of
Single-stage and Multi-stage Fermenters to Study
the Metabolism of Xenobiotic and Naturally
Occurring Molecules by Interacting Microbial
Associations." In: Microbiological Methods for
Environmental Biotechnology, Grainger, J. M. and
J. M. Lynch, Editors. Society for Applied
Bacteriology, Academic Press, Inc., Orlando, FL.
pp. 275-293.
Slater, J. H. and D. Lovatt, 1984. "Biodegradation
and the Significance of Microbial Communities."
Microbial. Degradation of Organic Compounds, D.
T. Gibson, Editor. Marcel Dekker, Inc., New York,
pp. 439-485.
Smolenski, W. J. and J. M. Suflita, 1987. "Biode-
gradation of Cresol Isomers in Anoxic Aquifers."
Appl. Environ. Microbiol., Vol. 53, pp. 710-716.
Suflita, J. M. and S. A. Gibson, 1985. "Biodegra-
dation of Haloaromatic Substrates in a Shallow
Anoxic Groundwater Aquifer." Proc. Second Intl.
Conf. Groundwater Quality Res., N. N. Durham
and A. E. Redelfs, Editors. Natl. Cntr. Ground-
water Res., Stillwater, OK, pp. 30-32.
Suflita, J. M. and G. D. Miller, 1985. "The Microbial
Metabolism of Chlorophenolic Compounds in
Groundwater Aquifers." Env. Toxicol. Chem., Vol.
4, pp. 751-758.
Updegraff, D. M., 1982. "Plugging and Penetration of
Petroleum Reservoir Rock by Microorganisms."
Proceedings of 1982 International Conference on
Microbial Enhancement of Oil Recovery, May 16-
21, Shangri-La, Afton, OK.
van Beek, C. G. E. M. and D. van der Kooij, 1962.
"Sulfate-Reducing Bacteria in Ground Water from
Clogging and Nonclogging Shallow Wells in the
Netherlands River Region." Groundwater, Vol. 20,
pp. 298-302.
Ventullo, R. M. and R. J. Larson, 1985. "Metabolic
Diversity and Activity of Heterotrophic Bacteria in
Ground Water." Env. Toxicol. Chem., Vol. 4, pp.
321-329.
Waksman, S. A., 1916. "Bacterial Numbers in Soil,
at Different Depths, and in Different Seasons of
the Year." Soil Science, Vol. 1, pp. 363-380.
Ward, T. E., 1985. "Characterizing the Aerobic and
Anaerobic Microbial Activities in Surface and
Subsurface Soils." Environ. Tox. Chem., Vol. 4, pp.
727-737.
Webster, J. J., et al., 1985. "Determination of
Microbial Numbers in Subsurface Environments."
Groundwater, Vol. 23, pp. 17-25.
White, D. C., et al., 1985. "Biochemical Measures of
the Biomass, Community Structure, and Meta-
bolic Activity of the Ground Water Microbiota."
In: Ground Water Quality, C. H. Ward, W. Giger,
and P. L. McCarty, Editors. John Wiley & Sons,
Inc., New York, pp. 307-329.
White, D. C., et al., 1983. "The Groundwater Aquifer
Microbiota: Biomass Community Structure and
Nutritional Status." Dev. Ind. Microbiol., Vol. 24,
pp. 201-211.
Wilson, J. T., et al., 1983. "Enumeration and Charac-
terization of Bacteria Indigenous to a Shallow
Water-table Aquifer." Groundwater, Vol. 21, pp.
134-142. ;
Wilson, J. T., et al,, 1985, "Influence of Microbial
. Adaption on the Fate of Organic Pollutants in
Ground Water." Env. Toxicol. Chem., Vol. 4, pp.
743-750. ..-',.,
Wilson, J. and M. J. Noonan, 1984. "Microbial
Activity in Model Aquifer Systems." In: Ground-
water Pollution Microbiology, Bitton, G., and C. P.
Gerba, Editors. John Wiley & Sons, New York, pp.
117-133.
Wilson, J. T., M. J. Noonan, and J. F.! McNabb, 1985.
"Biodegradation of Contaminants in the Sub-
surface." In: Ground Water Quality, C. H. Ward,
W. Giger, and P. L. McCarty, Editors. John Wiley
& Sons, Inc., New York, pp. 483- 498.
Wood, P. R., R. F. Lang, and I. L. Payan, 1985.
"Anaerobic Transformation, Transport and
Removal of Volatile Chlorinated Orgariics in
Ground Water." In: Ground Water Quality, C. H.
Ward, W. Giger, and P. L. McCarty; Editors. John
Wiley & Sons, Inc., New York, pp. 493-511.
Zoetman, B. C. J., 1985. "Overview of Contaminants
in Ground Water." In: Ground Water Quality, C. H.
Ward, W. Giger, and P. L. McCarty, Editors. John
Wiley & Sons, Inc., New York, pp. 27-38.
84
-------
CHAPTER 8
MICROBIOLOGICAL PRINCIPLES INFLUENCING THE
BIORESTORATION OF AQUIFERS
Joseph M. Suflita
Introduction
Ground-water pollution problems can be enormously
complex. Symptoms of aquifer contamination include
odor and taste problems and the occasional appear-
ance of a free product phase in wells. Besides the
health concerns associated with the latter type of
contamination, volatile and flammable fumes can
accumulate in buildings and result in explosions
and/or inhalation risks. Ground-water contamina-
tion also can conceivably impact surface water
supplies and lead to more obvious signs of pollution
Including color and odor problems, fish kills, and if
an immiscible chemical is involved, the formation of
seeps or slicks.
The purpose of this chapter is to briefly consider
various treatment options for cleaning up contam-
inated aquifers and to illustrate how biorestoration
techniques fit into the myriad of pollution mitigation
tools. This chapter also describes the types of consid-
erations that must be made prior to implementing a
field aquifer biorestoration program. An example of
spilled gasoline in an aquifer is chosen to illustrate
how basic microbiological and biochemical principles
meld into an overall aquifer treatment strategy. In
addition, guidelines are provided for the critical
evaluation of claims for aquifer restoration, with
suggestions for the types of information that might
be collected to support such claims. Particular
attention is paid to in situ biorestoration attempts
that rely on the inoculation of desirable micro-
organisms. Lastly, a perspective on bioreclamation
techniques is provided through a consideration of the
practical limitations of the technology.
Environmental Fate of
Contaminants and Treatment Options
At first glance, it may seem unusual to consider the
environmental fate of a contaminant together with
various options for the abatement of that contam-
inant. However, these topics are governed by the
same two phenomena - the transport characteristics
of a pollutant and the reaction of the contaminant
with the environment.
Table 8 generally depicts the fate of a contaminant in
a specific environment in terms of movement, reten-
tion, and reaction processes as a function of the
properties of both the environment and the contam-
inant. The fate of a contaminant is largely a function
of the chemistry of the pollutant in its environment.
Similarly, the reaction mechanisms that a pollutant
may undergo (i.e., hydrolysis, precipitation, oxida-
tion/reduction, etc.) are a function of that chemical
and the existing environmental conditions. These
various processes are interrelated (Table 8). For
instance, a chemical may be inherently susceptible
to microbial attack, but is deposited in an environ-
ment where the production of low molecular weight
acids from the microbial metabolism of other forms of
organic matter results in a decrease in overall pH.
The immediate pH conditions may prevent the
continued metabolism of that chemical. The inter-
relationships between environmental conditions,
chemical structure, and the physiology of micro-
organisms are described in Chapter 7.
Treatment options for contaminated aquifer remedi-
ation are governed by the same basic characteristics
85
-------
(Table 9). Transport and reactivity properties of a
contaminant can be envisioned as a continuum from
low to high for each characteristic, and can help
decision-makers who are considering an aquifer
treatment option. For example, if a contaminant does
not transport well and tends to be a recalcitrant
molecule, containment may be the most likely
treatment option. However, if the pollutant is subject
to desirable biotransformation reactions and reactiv-
ity is high, then a bioremediation approach to
contaminated aquifers should at least be considered.
The simple analysis in Table 9 suggests that restora-
tion efforts centered on biodegradation strategies are
not applicable in all circumstances.
Table 8. Summary of the Mechanisms Influencing the
Fate of Contaminants in Environments
Process
Movement
Retention
Reaction
Environmental
Conditions
Water Flow Rate
Formation
Permeability
Water Motion
Gravity
Surface Tension
Soil/Sediment
Organic Matter
Content
Sorptive Capacity
pH
Redox Status
Microbial
Communities
Contaminant
Amount of Material
Physical State
Solubility
Viscosity
Type Solubility
Ionic Character
Chemical
Transformation
Biodegradability
Table 9.
Reactivity
Treatment Options as a Function of the Trans-
port and Reactivity Characteristics of the
Contaminant
Transport
High
Low
High
Low
Extraction and/or
Biodegradation
Extraction
Biodegradation
Containment
In fact, all remediation technologies have inherent
limitations that must be kept in mind when evalu-
ating a variety of treatment options. For instance,
containment techniques generally rely on physical or
hydrological barriers to keep a contaminant from
spreading underground. Typical containment meth-
ods involve slurry walls, clay caps, interceptor
trenches, or hydraulic barriers. These techniques can
be costly, plus there is no guarantee of their long-
term effectiveness and, most importantly, they do
not really address the pollution problem.
Similarly, extraction-based remediation efforts also
have disadvantages. As noted in Table 9, the technol-
ogy tends to be used on substances that are easily
transported in the subsurface. The simplest extrac-
tion approach is to excavate the contaminated soils
and sediments in the problem area. However, this
often is not feasible because of the depth of the
problem area or because excavation activities could
undermine the foundation of roads, buildings, or
other structures.
The pumping of ground water is another extraction
technique. Ground-water pumping technologies
generally are combined with some sort of surface
treatment like air stripping, carbon adsorption,
biological or chemical reaction, or simple discharge
to other locations. Often such extraction techniques
tend to treat only a limited amount of the total
contaminant problem; that is, only contaminants
dissolved in the water phase are removed from the
aquifer. Many decades often are needed to extract
even mildly adsorbing chemical pollutants.
One of the major disadvantages of extraction
technologies is that the contaminants often are
merely transferred from one location to another. The
problems associated with trans-media pollution of
different environmental compartments continue to
gain increasing attention among the regulatory
community. Other methods, such as soil venting,
tend to address volatile contaminants present in the
soil atmosphere, yet again, only address a portion of
the problem.
Extraction often tends to be the treatment of choice
because it is straightforward, understandable, and
predictable and generally has small yearly invest-
ment costs. However, this approach ignores the long-
term commitment to the extraction process, partic-
ularly for sorbed contaminants and the maintenance
involved in the treatment operation.
Reaction technologies include both chemical and
biological methods of converting contaminants to
more environmentally acceptable forms. Generally,
chemical transformations convert pollutants
through hydrolysis or initial oxidations. Such
reactions also can be catalyzed by microorganisms,
but they often are the rate-limiting step in biological
treatment scenarios. On the other hand, biological
treatment processes have the potential to completely
mineralize contaminants and convert them to
innocuous substances like carbon dioxide, water, and
cell material. Reaction processes, when successful,
tend to be relatively short-term solutions to pollution
problems and are associated with high operation
costs. The treatment chemicals and/or microbial
nutrients must be transported to the contaminated
86
-------
portion of the aquifer so that both the dissolved and
adsorbed phases of the contaminant are attacked.
Aquifer Remediation Considerations
Many considerations must be made before seriously
considering an aquifer bioremediation approach
(Lee, et al., 1988). The first consideration is
determining the contaminant type, the phases in
which it exists, the solubility of the material, and its
inherent susceptibility to biodegradation. The
microbial ecology scientific literature is an excellent
source for the latter consideration. Biodegradation
information resulting from the study of a variety of
environments indicates whether the contaminant
can be degraded and, if so, under what ecological
conditions. This information does not insure that a
contaminant will be degraded in all instances since a
variety of site-specific characteristics can interplay
to influence the success of bioremediation efforts.
The scientific literature also can provide some as-
sessment of the likely pathways of degradation,
which in turn allows for an evaluation of the
consequences of biorestoration. Disappearance is
only one aspect of biodegradation since a contam-
inant can be converted to a series of metabolites that
also are of environmental concern. For instance, the
reductive removal of tetrachloroethylene (TeCE)
under anaerobic conditions leads to a series of
dehalogenated intermediates. TeCE's halogens are
removed and replaced by protons in a series of
sequential steps. However, the rate of reductive
dehalogenation tends to decrease as fewer and fewer
halogens remain on the ethylene moiety. Conse-
quently, vinyl chloride accumulates and, from a
regulatory viewpoint, causes greater concern than
the parent contaminant. Thus, biodegradation path-
way information can be critical for the design of
bioremediation efforts.
Successful implementation of biorestoration efforts
also depends on determining site characteristics, in
particular, the site-specific hydrogeological vari-
ables. Factors such as thickness of the vadose zone,
its permeability, geologic complexity, and organic
matter content can impact biorestoration programs.
Saturated zone characteristics such as the type and
composition of an aquifer, its permeability, thick-
ness, interconnection to other aquifers, location to
discharge areas, magnitude of water table fluctu-
ations, and ground-water flow rates all influence
aquifer bioremediation planning and decision-
making. To date, bioremediation has been attempted
in aquifers possessing a variety of flow charac-
teristics including pumping rates ranging from 25 to
380 L per minute, and flow rates from 0.6 to 800 m
per year or hydraulic conductivities of 10-5 to 10-3 cm
per second. Generally, bioremediation efforts center
around more permeable aquifer systems where the
movement of ground water can be more successfully
controlled.
Another consideration in biorestoration is the
removal of free product. This is extremely important
since many substances, while suitable nutrients for
microbial growth when present at low concen-
trations, are inhibitory at high concentrations. Such
concentration effects can easily be observed with
substances ranging from simple sugars to gasoline.
Consequently, in a biorestoration effort, it is impor-
tant to remove as much free product as possible by
physical or chemical means and use biological
methods to treat the remainder.
The next biorestoration consideration is whether the
system design will be above or below the ground.
Again, this decision is influenced by the ease of
transporting the contaminant from its location to the
area where the treatment will take place. Since
many contaminants are in multiple phases in the
subsurface, in situ treatment strategies often are the
best option. In in situ treatment efforts, nutrients are
transported to the requisite microorganisms in an
effort to create the correct environmental conditions
for microbial proliferation.
A critical component in biorestoration is a laboratory
evaluation of the site pollutants' susceptibility to
biodegradation. The appropriate use of microcosms
in rigorously defined and controlled experiments
allows the investigator to evaluate biodegradation
under conditions that are more environmentally
realistic. The microcosm approach allows the investi-
gator to incorporate numerous site-specific variables
without a detailed knowledge of what these variables
actually are. For instance, scientific literature may
indicate that a particular pollutant is subject to bio-
degradation, but this may not be confirmed by
microcosm studies due to the absence of an essential
microbial nutrient or the presence of an inhibitory
substance at the site. Some of these limitations may
be relatively easy to overcome, while others may
prove more difficult.
If initial microcosm evaluations are unsuccessful,
subsequent studies should attempt to stimulate
biodegradation of the problem contaminants using
nutrients or adjusting other variables. If biodegra-
dation in microcosms cannot be successfully stimu-
lated, the chances of success in the field are minimal
but if initial microcosm evaluations prove successful,
the chances for success in the field are significantly
improved. These preliminary laboratory investiga-
tions not only allow an investigator to design a
biorestoration program, but also to optimize the
treatment strategy.
87
-------
As a final consideration, some plan should be devised
to monitor the progress of a biostimulation activity.
A more detailed discussion of the types of informa-
tion needed for monitoring are presented later in this
chapter.
From a bioremediation viewpoint, the ideal site has:
(1) a homogeneous and permeable aquifer; (2) a con-
taminant originating from a single point source; (3) a
low ground-water gradient; (4) no free product; (5) no
soil contamination; and (6) an easily degraded, ex-
tracted, or immobilized contaminant. Obviously, the
above combination of characteristics describes very
few sites. The following sections will attempt to show
how information on the chemistry, microbiology, and
biochemistry of a contaminant can form a bio-
restoration strategy.
Pathways of Hydrocarbon Metabolism
One of the most frequently encountered aquifer con-
taminants is gasoline, which is largely a mixture of
hydrocarbons including three major chemical
classes: the alkanes; the cycloalkanes; and aromatic
chemicals (Table 10). Table 10 illustrates that
greater than 73 percent of several gasolines are
composed of these three hydrocarbon classes. Gaso-
line hydrocarbons within each class possess a variety
of substituents and numerous substitution patterns
that ultimately influence the susceptibility of the
resulting structures to biodegradation.
Table 10. Major Chemical Components (Percent of Total)
of the Gasoline Fraction of Several Petroleums
(Adapted from Perry, 1984)
Hydrocarbon
Class <
Alkanes
Cycloalkanes
Aroma tics
Texas
16.8
47.1
19.5
Gasoline from
California
18.0
55.5
10.2
Louisiana
24.5
38.4
15.6
Alkanes exist in gasoline and many gasoline
components contain an alkane moiety in their carbon
skeleton. An enormous amount of scientific litera-
ture exists on the types of alkanes that are subject to
microbial attack and the diversity of microorganisms
involved in alkane metabolism as detailed by Singer
and Finnerty (1984). An extremely wide range of
alkanes are susceptible to biodegradation, but some
microorganisms preferentially utilize only short
chain molecules while others metabolize only longer
chain structures.
Microorganisms can initiate the aerobic degradation
of alkanes through mechanisms that involve mono-,
di-, or subterminal oxidation of the molecule. Prob-
ably the most frequently reported degradative route
involves the oxidation of a terminal methyl group to
convert the alkane to an alcohol of equivalent chain
length (Figure 76). This alcohol probably undergoes
a series of dehydrogenation steps to form the
corresponding aldehyde and eventually a fatty acid
(Singer and Finnerty, 1984).
Once formed, the fatty acid is further metabolized by
{5-oxidation. This process liberates a two-carbon
fragment as acetylcoenzyme A, a central metabolic
intermediate. Two carbon fragments are then con-
tinuously cleaved from the resulting shorter fatty
acid until the entire molecule is degraded. Alkanes
with odd numbers of carbon atoms ultimately yield a
molecule of another central metabolite - propionyl
coenzyme A. The significant aspect of Figure 76 is
that the initial oxidation of an alkane molecule
involves the incorporation of molecular oxygen. This
basic theme also is pertinent to the other major
classes of gasoline hydrocarbons.
CH3 - (CH2)n
+ 02.2Hl
P^-
CH3 - (CH2)n - CH3 Alkane
/ Methyl
H20 ! Group
CH3 - (CH2)n - CH2OH Alcohol
< ~2H
CH3 — (CH2)n — CHO Aldehyde
H20 -v -2H
Ti r • J
CH3 — (CH2)n — COOH Fatty Acid
I
0 -Oxidation
CO2
Figure 76.
t
The initial steps in the aerobic microbial
metabolism of alkane hydrocarbons.
It is more difficult to enrich for microorganisms that
are able to use cycloalkanes as a sole source of carbon
and energy. This appears to be widespread in nature
and is discussed by Trudgill (1984) and Perry (1984).
It may be that microorganisms interact frequently in
commensalistic relationships based on the cometab-
olism of these substrates (cometabolism will be dis-
cussed later in this chapter). However, it is clear that
individual or mixtures of microorganisms can metab-
olize cycloalkanes in the manner shown for cyclo-
hexane (Figure 77).
88
-------
O-*
2H
Cyclohexane
H2O
-2H
O2 + 2H
Cyclohexanol
Cyclohexanone
H20
COOH
CH2
0-Caprolactone
CH2
CH2
|3-Oxidation
•+• CO2 + H2O
Figure 77.
CH2
I
COOH
Adipic Acid
The Initial steps in the aerobic microbial metabolism of alicyclic hydrocarbons.
As shown in Figure 77, cyclohexane hydroxylation
by a microbially produced monooxygenase leads to
the formation of an alicylic alcohol. Subsequent
dehydrogenation of the alcohol forms a ketone and
further oxidation of the ketone results in the
formation of a lactone ring structure. The lactone is a
suitable substrate for ring opening and is eventually
converted to a dicarboxylic acid which, in turn, is
subject to oxidation. Molecular oxygen participates
in the biodegradation pathway, but, in this case, is
involved in two separate steps.
Aromatic compounds of gasoline also are subject to
microbial attack by many different types of bacteria
and fungi. The aromatic hydrocarbons are aero-
bically metabolized by bacteria to dihydroxylated
compounds through cis-dihydrodiol intermediates.
Figure 78 shows how, in the case of benzene, a
bacterial dioxygenase incorporates both atoms of
molecular oxygen to form cis-benzene dihydrodiol
which is subsequently dehydrogenated to result in
the formation of catechol. Catechol is then a suitable
substrate for ring cleavage. Other dioxygenases open
the ring most often via an 'ortho' or 'meta' cleavage
route. The subsequent intermediates produced by
these pathways eventually enter central metabolic
reaction sequences of the bacterial cell. The aerobic
catabolism of homocyclic aromatic compounds is
reviewed by Gibson and Subramanian (1984) and by
Bayly and Barbour (1984).
Oxygen plays a critical role in the metabolism of
aromatic hydrocarbons. All of the metabolic path-
ways discussed above require oxygen as a coreactant.
In addition, the organisms catalyzing these biocon-
versions use oxygen as a terminal electron acceptor
(described in Chapter 7). Hydrocarbon metabolism
puts a large demand on oxygen resources; therefore,
plans for biorestoration activities should consider
how this oxygen demand will be supplied.
In addition to oxygen, other potential limiting
nutrients also must be supplied in suitable form for
the microorganisms to proliferate at the expense of
the hydrocarbons. Along with a suitable environ-
ment, microorganisms need nitrogen, phosphorus,
sulfur, and trace elements and without the two
former nutrients, hydrocarbon metabolism may be
limited even when oxygen supplies are adequate.
With the proper nutrients, the microorganisms can
convert gasoline hydrocarbons to the environ-
mentally innocuous products of carbon dioxide,
water, and additional cell material.
Once the chemical and physical requirements for
microbial growth and the predominant metabolic
pathways are known, attempts can be made to
stimulate the biodegradation of gasoline by
superimposing the correct nutrients in situ. Figure
79A represents an aquifer contaminated with gaso-
line hydrocarbons. Free product already has been
removed from this schematic site, but both water and
soil contamination still exist. Preliminary testing
shows that hydrocarbon-degrading microorganisms
are present and active at the site and that these
organisms can be stimulated to increase hydrocarbon
89
-------
Benzene
cis-Benzene Oihydrodiol
Muconic Semialdehyde
COOH
COOH
cis, cis-Muconic Aci<
_ Figure 78.
Mineralization
CO2 + H2O
The initial steps in the aerobic microbial metabolism of aromatic hydrocarbons.
biodegradation. Figure 79B depicts the attempt to
gain hydrogeologic control over the contaminated
area through the insertion of a series of injection and
recovery wells. The goal is to recirculate the ground
water in an effort to create an underground reaction
chamber. The figure shows that the pumping and
recirculation also change the water table. Ideally,
the 'reaction chamber1 will encompass the contam-
inated area; however, this often is not possible due to
the magnitude of the contamination problem and the
permeability limits of the aquifer. In the latter case,
the contamination problem can be approached in
stages that are governed largely by the efficiency of
water recirculation at the site.
Nutrients are then added to the recirculation system
and transported with the injection water into the
aquifer (Figure 79C) where the indigenous micro-
organisms will start to proliferate (the bioactive
area). The efficiency of the nutrient transport will
then dominate the success of the remediation effort.
A variety of mechanisms for supplying oxygen to the
requisite microorganisms is considered in the review
byLee,etal.(1988).
Initially, most of the bioactive area probably will be
centered in areas adjacent to the injection well. As
the electron donors in that area are depleted, the
major bioactivity shifts to other areas of the contam-
inant plume where additional hydrocarbon and,
thus, electron donors exist, and the requisite nutri-
ents penetrate the plume (Figure 79D). This process
continues until the site is considered remediated
(Figure 79E). At this point, the nutrients are
removed from the recirculation stream. The aquifer
tends to exhibit homeostatic controls (described
below) and the microflora that grew in response to
the hydrocarbon and nutrient input return to the
levels originally present before the contamination
incident.
Bioreclamation of gasoline in aquifers has an
excellent chance of success because of the wealth of
scientific information that forms the basis of this
applied technology. Researchers have studied hydro-
carbon metabolism by microorganisms for many
decades and learned that gasoline hydrocarbons are
xenobiotic substrates only in the sense that they
occur in various environmental media at higher than
acceptable concentrations. Most hydrocarbons in
gasoline are natural substrates and microbial com-
munities have evolved mechanisms for their degra-
dation. These hydrocarbon-degrading microorgan-
isms are not unusual and tend to be ubiquitously
distributed in many and probably most natural
environments. Scientific investigations also have
helped elucidate the nutrient requirements neces-
sary for microbial growth and proliferation at the
expense of hydrocarbon substrates so that remedia-
tion activities are relatively easy to construct, test in
microcosms, and extrapolate to the field.
90
-------
^-Contominoted oreo
"-.Original
Water Table
Recycled
Ground
Woter-r
New
Water Table
"-Original
/ Water Table
Recycled
Ground
Water
Mutrient Flow
] Bioactive Area
New
Water Table
Original
Water Table
New
Water Table
Original
Water Table
New
Water Table
Original
Water Table
\ Contaminant
Figure 79. Schematic illustration of the various steps
(A-E) in an aquifer biorestoration program
(see text for description).
A similar information base does not exist for more
exotic substrates; therefore, microcosm research may
be necessary in order to design effective remediation
processes. Principles gleaned from the study of
hydrocarbon degradation may or may not be applica-
ble to other contaminants.
Critical Evaluation of Biorestoration
Claims have been made that a variety of hydro-
carbons, solvents, and other contaminants were
treated by in situ bioremediation (Table 11) (Lee, et
al., 1988). However, these claims generally lack firm
scientific evidence that biorestoration actually was
responsible for the removal of the substances. Many
claims for in situ biorestoration tend to appear in
publications that were not peer reviewed, therefore,
legitimate questions persist as to the effectiveness of
biological treatment programs for various contami-
nant types. Consequently, clients or regulators have
difficulties evaluating such claims. Since biorestora-
tion is still an emerging technology, practitioners
must gather the necessary evidence for prospective
clients and regulators to critically interpret the
chances of success for such programs.
Table 11. Contaminants Treated by In Situ Bioremedi-
ation (Adapted from Lee, et al., 1988)
Contaminant Type
Representatives
Hydrocarbons
Solvents
Other Compounds
gasoline
mineral oil
aliphatic plasticizers
methyl chloride
n-butanol
acetone
ethylene glycol
isopropanol
tetrahydrofuran
chloroform
dimethyl aniline
In remedial activities, practitioners should carefully
document the reduction in substrate concentration(s)
as a result of their efforts. Implicit in this suggestion
is reliable information on mass balances. While it is
often impossible to estimate the quantity of material
lost in an accident, sometimes such estimates are
feasible. For instance, logs are kept of inputs and
removals of chemical substances from storage facil-
ities. This type of information may prove valuable in
estimating the maximum amount of contaminant
involved in a pollution incident. Alternately, the
remedial efforts can be gauged relative to the initial
amount of contaminant present in the aquifer. This
approach requires good definition of the plume, an
appreciation for where the contaminant exists, and
reliable and specific measurement methods. If biore-
mediation is successful, contaminant levels should
begin to fall in areas that receive the treatment and
remain relatively unchanged in areas that do not.
Bioremediation efforts often seek to increase the
activity of microorganisms in an aquifer. The
confidence in bioremediation technologies would be
greater if the increase in microbial numbers and/or
activities were quantitated relative to: (1) plume
areas prior to any treatment; (2) areas within the
plume that did not receive the treatment; or (3)
control areas outside the plume. The latter will give
some indication of background levels of micro-
organisms and allow comparisons to be made before,
during, and after a remediation program.
91
-------
Often, the production of microbial catabolites can be
measured in areas that receive treatment but not in
contaminated areas left untreated. For instance, the
production of a variety of lesser halogenated PCB
congeners in river sediments contaminated with
Aroclors was previously noted (Brown, et al., 1987;
Brown, et al., 1984). Many of the congeners produced
did not comprise a significant portion of the original
contaminant Aroclors. Their production was sug-
gested as field evidence that the original PCB
materials were metabolized by anaerobic micro-
organisms and reductively dehalogenated (Brown, et
al., 1987; Brown, et al., 1984). Laboratory evidence
in microcosm experiments designed to test this
hypothesis confirmed that suspicion (Quensen, et al.,
1988). Similarly, the production of metabolites like
hydrogen sulfide or methane often are testament to
the metabolic activities of microorganisms. If the
degradation pathways of particular contaminants
are known prior to a remedial effort, it is sometimes
possible to specifically assay for the production of
likely catabolites.
Many times, biorestoration programs rely on the
contaminant itself to supply the electron donors
necessary for microbial growth and proliferation.
One of the objectives of the remediation is to supply
the necessary electron acceptor. The consumption of
added terminal electron acceptors could possibly be
measured. For instance, if oxygen is added to help
stimulate hydrocarbon metabolism in aquifers, its
depletion in the treatment area should be relatively
easy to measure.
Finally, rapid microbial biodegradation typically
seems to be preceded by a variable length of time
where little or no activity is measured. This length of
time is referred to as the adaptation or acclimation
period (Linkfield, Suflita, and Tiedje, 1989). So far,
adaptation seems to be a unique biological response.
Observation of this type of phenomenon in response
to a biorestoration effort can be taken as presump-
tive evidence that microbial action is operative.
It is important to emphasize that all of the
techniques and measurements suggested above must
be made relative to appropriate controls. For some
techniques, that may mean the assay of untreated
areas or areas outside the plume. Many clients may
not be willing to invest the necessary time and
financial resources to gather this level of informa-
tion. Even with controls, the evidence garnered in
the above fashion tends to be largely circumstantial.
However, the stronger the evidence, the greater the
degree of confidence in the technology.
Lag, Adaptation, or Acclimation Periods
In biodegradation studies, a period of time often is
observed where very little substrate is turned over
and correspondingly little product is formed. This
phase of 'metabolism' is potentially more environ-
mentally significant than other phases because of its
possible effects on the level of the exposure of
humans and ecosystems to specific pollutants. The
lag, adaptation, or acclimation phase can be one of
the most frustrating portions of bioremediation
programs, since, despite all efforts, virtually nothing
seems to be happening to the problem contaminant.
For example, Figure 80A«'shows the reductive
dehalogenation of 2-bromobenzoate in anoxic sedi-
ment microcosms. The substrate lags for several
weeks after which it is rapidly metabolized. The
production of an intermediate catabolite, benzoic
acid, occurs before its rapid degradation. Ultimately,
the substrate and intermediate are converted to the
gaseous endproducts, methane and carbon dioxide.
At this point, the microorganisms are considered
acclimated to the degradation of the parent substrate
and subsequent substrate additions will be degraded
without a lag. This phenomenon is graphically
illustrated for the reductive dehalogenation of a
related substrate 3-iodobenzoate (Figure SOB). In
this case, the second addition of 3-iodobenzoate is
presented as time zero. Note the immediate con-
sumption of the substrate, the much lesser accumu-
lation of benzoic acid, and the more rapid conversion
of the parent molecule to gaseous endproducts.
In the examples given in Figure 80A and B,
acclimation might reasonably be expected since the
microorganisms were able to mineralize the parent
substrates. Some of the carbon and energy gained
from the metabolism of those substrates was
presumably used for the proliferation of the cata-
lyzing microbial communities. If the concentration of
the requisite microorganisms was initially very low,
the lag period could conceivably be a function of the
time required for these organisms to grow to
sufficient density to effect some significant amount
of substrate depletion relative to the large back-
ground added as the amendment. However, acclima-
tion also is observed for substrates that are not
mineralized. As seen in Figure 80C, the substrate 4-
amino-3,5-dichlorobenzoate can be reductively de-
halogenated to form the monohalogenated product
following a three-week lag period. Even though the
parent substrate is not mineralized and does not
supply carbon for microbial growth, additions of the
parent substrate following the removal of the initial
92
-------
,4-Amino-3,5
dichloro-
banzoate
4-Amino-3-chloro
btnjoalt
100
oc.
o
50°
CE
O
a
o
cc
o_
21
24 27 30 33
J 0
INCUBATION TIME (DAYS)
Figure 80.
Patterns of anaerobic reductive dehalogenation of halobenzoates by sediment microorganisms: (a) degradation
of 2-bromobenzoate in fresh sediment microcosms; (b) degradation of 3-iodobenzoate by sediment micro-
organism previously acclimated to 3-iodobenzoate degradation; (c) degradation of 4- amino-3,5-dichlorobenzoate
by fresh sediment and the accumulation of the monodehalogenated endproduct (Horowitz, et al., 1983).
amendment are metabolized without an additional
lag period.
Examining and understanding the factors which
influence the lag period can be difficult, but may
ultimately lead to biorestoration scenarios with
controlled, reduced, or even eliminated adaptation
times. The requirement for the growth of the
requisite microorganism most often is touted as the
operational reason for lag periods. While undoubt-
edly true, the example above illustrates that there
may be other reasons for the delay in biodegradation.
The structure of the chemical itself is known to
influence the rate of biodegradation. Figure 81
compares the adaptation period for a variety of
halobenzoates in methanogenic sediment micro-
cosms. Note that all the substrates were added at the
same starting concentration and that they all
possessed halogen substitutions at one or both meta
(3 or 5) positions. When degraded, all of the
substrates were metabolized via reductive dehalo-
genation reactions. The various substrates were
degraded in a specific order; that is, 3-bromobenzoate
(3-Br-Bz) degraded before 3-iodobenzoate (3-I-Bz)
which in turn degraded faster than 3,5-dichloro-
benzoate (3,5-diCl-Bz), etc. Therefore, chemical
structure very definitely influences the length of the
lag period, and, perhaps more significantly, the
length of the lag periods are relatively reproducible.
The insert table in Figure 81 illustrates that a repeat
experiment not only gives the same relative order of
degradation, but also approximately the same length
of the lag period. The lag period does not appear to
correlate with whether the substrate is mineralized.
This result implies that a specific physiological or
chemical basis exists for the characteristic lag
periods.
Other experiments show that concentration can
markedly influence the length of the lag period.
Figures 82 and 83 compare the reductive dehalo-
genation of several halobenzoates at various concen-
trations and illustrate a number of characteristic
patterns. In the case of the anaerobic biodegradation
of both 3,5-dichloro- and 3-chlorobenzoate at sub-
strate concentrations ranging from 20-800 uM
(Figure 82), the dichloro-substrate exhibits a
characteristic lag time prior to rapid biodegradation
regardless of the substrate concentration range.
However, the length of the lag period associated with
the lower concentrations of monochlorobenzoate is
much shorter than those observed with higher
substrate concentrations. This perhaps is not sur-
prising since the benzoates are known bacteriostatic
agents and higher concentrations tend to inhibit
microbial activity.
A similar situation also can be observed in the case of
low vs. high concentrations of 3-fluorobenzoate
(Figure 83). However, the opposite result is observed
when the substrate is changed to 4-amino-3,5-
dichlorobenzoate. With this substrate, concentra-
tions 3: 40 uM were degraded with a characteristic
lag period. However, concentrations of 20 uM ex-
hibited lag periods in excess of one year (data not
93
-------
4-NH2-3,5-diCI-BZ
3,5-diCI-BZ
ADAPTATION PERIOD (DAYS)"
EXPT 3-Br 3-1 3,5-diCI 4-NH2- 3,5 di Cl 3'CI 3~F
150-170 >I70
125-148 >365
28 29-35
60
TIME (DAYS)
200
240
Figure 81.
Length and reproducibility of the adaptation period prior to the reductive dehalogenation of several
halobenzoates at initial concentration of 800 urn (Linkfield, et al., 1989).
completely shown). If the concentrations of this
substrate were only doubled, biodegradation would
proceed in typical fashion and reach levels far below
20 uM. From these examples, one can see that when
substrate concentrations are either too high or too
low, biodegradation activity can be adversely
affected and increased lag periods could result.
Potential Biostimulation Approaches
Many other factors besides the necessity for
microbial growth, chemical structure, and substrate
concentration are known to influence the length of
the lag period. These factors include the need to
deplete competing substrates and exchange genetic
information and the lack of required nutrients. An
important issue is whether biodegradation can be
stimulated so that the lag or adaptation period can
be overcome as quickly as possible. This can happen,
provided that the factors controlling the lag and
adaptation periods are understood. Although some
biostimulation approaches were discussed earlier,
three additional approaches will be considered: cross
adaptation; analog enrichment; and biomass enrich-
ment.
The principle of cross adaptation involves the
addition of a readily degradable substrate to help
bring about the rapid biotransformation of more
recalcitrant molecules. An example of this phenom-
enon is illustrated in Table 12 and the study
described below.
Anoxic sediment microcosms were exposed to one of
two haloaromatic substrates, the relatively easily
degradable 3-bromobenzoate and the more recalci-
trant 4-amino-3,5-dichlorobenzoate. As indicated in
the table, the former substrate would start to de-
grade in as little as a few days, whereas the latter
substrate normally took about eight weeks. In both
cases, subsequent additions of the same substrates to
the microcosms were degraded without an observ-
able lag period. Complete degradation of 4-amino-
3,5-dichlorobenzoate reamendments took two to
three weeks while 3-bromobenzoate additions took
less than one week.
If other halobenzoates were added to adapted
sediment instead of additional parent substrate, a
variety of responses were observed. When the
sediment microflora was adapted to the degradation
of a relatively labile substrate (3-bromobenzoate), it
94
-------
800 pM 3-GI-BZ
3,5-diCl-BZ
400uM 3,5-diCl-BZ
80 120 160 200
TIME (DAYS)
240
Figure 82.
Effect of substrate concentration on the length of the adaptation period prior to the reductive dehalogenation of
3,5-dichloro- and 3-chlorobenzoate (Linkfield, et al., 1989)
also was cross-adapted and capable of rapid
degradation of more recalcitrant materials like 4-
amino-3,5-dichlorobenzoate and 3,5-dichloroben-
zoate. However, this technique did not result in a
significant improvement in the degradation of other
substrates like 3-iodobenzoate or 3-chlorobenzoate.
Sediment microflora capable of degrading either of
the two starting substrates exhibited the same
apparent substrate specificity, while similar experi-
ments with 3-iodobenzoate-adapted organisms could
only degrade subsequent additions of the parent
substrate.
A technique that may be related to cross adaptation
is referred to as analog enrichment and is illustrated
by the experiments of You and Bartha (1982). Their
objective was to stimulate the mineralization of 14C-
labelled dichloroaniline by soil microorganisms.
These investigators added aniline as a structural
analog of the halogenated pollutant to soil experi-
ments. They found that the amount of mineralization
of dichloroaniline (determined as the amount of
14COa) was proportional to the amount of aniline
added as a substrate analog to soil (Figure 84). Pre-
sumably, the dichloroaniline was not a particularly
good inducer of the requisite enzymatic machinery
among the microbial communities in soil. The addi-
tion of the analog may have derepressed the enzymes
Table 12. Cross-adaptation of Anaerobic Microorganisms
to the Reductive Dehalogenation of a Variety of
Halobenzoates (Adapted from Horowitz, et at.,
1983)
Time (wk) for Complete
Degradation in Sediment
Substrate Tested
for Cross-
Adaptation
rT)
O
A
OL
Adaptation
Time
(wk)
3-8
0.5-4
2-3
2-3
32-40
Adapted to:
d j&
2-3 2-3
<1 <1
2-3 . 2-3
<1 <1
Lag Lag
responsible for aniline biotransformation and these
enzymes also might have recognized the halogenated
aniline as a suitable substrate. This study also
showed an increase in the microbial mineralization
of humus-bound dichloroaniline as a function of
analog enrichment with aniline; humus-bound
residues generally are considered to be much more
recalcitrant than unbound residues.
Finally, the results of Wilson and Wilson (1985) may
illustrate an enrichment of desirable microorgan-
95
-------
40OpM 4-NH2-3,5-diCI-BZ
160
20O 240
TIME (DAYS)
Figure 83. Differences in the adaptation period as a
function of substrate concentration for two
halobenzoates at 20 and 400 pm initial
concentrations (Linkfield, et al., 1989).
75
Days
Figure 84. Effect of aniline on the mineralization of 3,4-
dichloroaniline in soil: (a) 1.8 mg of aniline
added per gram; (b) 0.4 mg of aniline added
per gram; (c) no aniline added; (d) HgCI2-
poisoned control (adapted from You and
Bartha, 1982).
isms in a complex microcosm. These authors found
that trichloroethylene (TCE) could be removed from
soil microcosms that were treated with a combina-
tion of methane and air. However, the TCE was not
removed (other than due to abiotic loss) in soil
microcosms that were not so treated. When the
methane- and air-amended microcosm was treated
with a bactericidal substance, the removal of TCE
stopped. The authors speculated that the micro-
organisms responsible for the biotransformation of
TCE were a unique group collectively called the
methane-oxidizing bacteria. These organisms pro-
duce a powerful oxygen-requiring enzyme called a
methane monooxygenase that is responsible for the
initial bioconversion of methane. However, the
enzyme has a broad substrate specificity and also can
oxidize other substrates including TCE. Other types
of bacteria not involved in methane metabolism can
act similarly (Nelson, et al., 1987; Nelson, et al
1986).
Presumably, the addition of methane and air to the
soil columns stimulated the proliferation of one or
several populations of methane-oxidizing micro-
organisms. While normally present in soil, the
relative numbers of these organisms represented a
greater proportion of the total microbial community
receiving the methane and air treatment. While
oxidizing methane, the organisms also cometabolized
TCE. That is, the organisms growing on a particular
substrate gratuitously oxidized a second substrate
which they were unable to use as a source of carbon
and energy for microbial growth. The cometabolized
substrate is not usually assimilated by the first
organism, but the oxidation products are then avail-
able for other organisms. This type of microbial
interaction forms the basis of many different types of
commensalistic relationships between microbial
populations. In fact, many different substrates are
known to be cometabolized, ranging from simple
short chain aliphatic hydrocarbons to complex halo-
genated pesticides (Horvath, 1972).
Bioremediation with Microbial Inoculants
The use of specialized microbial cultures for
bioremediation of contaminated aquifers is contro-
versial (these cultures may or may not involve the
use of genetically engineered organisms). Conclusive
proof that the use of such inoculants constitutes a
wise management practice has not been shown. Lee,
et al. (1988) wrote that the role of inoculants in
ground water cleanup efforts generally can not be
conclusively determined because adequate controls
many times are not employed in most experimental
designs. Fundamentally, questions concerning the
addition of microorganisms to aquifers are not
different from those surrounding the use of microbial
inoculants in soil to elicit some desirable response. In
the late 1970s, manufacturers captured the imag-
ination of the popular press with claims for such
products. These ranged from an assurance for
increased crop yields to the ability of such products to
make "depleted soils come alive." Generally, these
inoculants or microbial fertilizers have failed to live
up to their claims, which should not be surprising to
microbial ecologists. An excellent consideration of
96
-------
the relationships between microbial inoculants and
microbial ecological principles is given by Miller
(1979).
";' 4#v
Scientific literature shows examples of microbial
inoculation attempts in soil. The early work of
Katznelson (1940a; 1940b) can be considered
pioneering. Table 13 summarizes some of his
experiments on the survival of several bacteria,
actinomycetes, and fungi upon reinoculation of these
organisms into manured or manured and limed soil.
In general, these organisms (as well as others) died
back very rapidly upon inoculation.
A spore-forming Bacillus species was apparently
able to maintain its numbers under some conditions.
Eventually, none of the organisms could be detected
in numbers significantly above their baseline levels
in soil. Similarly, the review of Miller (1979)
summarizes the findings of Van Donsel, et al. (1967)
on the survivability of fecal coliforms and fecal
streptococci introduced to soil at various times of the
year (Figure 85). In both winter and summer, the
fecal bacterial numbers were reduced by several
orders of magnitude. However, the decreased
temperatures of winter allowed the fecal organisms
to survive for a longer time before they were
eventually eliminated.
Implicit in Figure 85 is the role of abiotic factors like
temperature on the mechanisms for elimination of
non-indigenous or foreign microorganisms inocu-
lated in soil. In fact, a variety of abiotic factors may
act alone or collectively to inhibit the survival of
inoculant microorganisms. These factors are those
known to influence microorganisms in general and
include pH, temperature, salinity, water content,
and osmotic or hydrostatic pressure among others.
In addition to abiotic factors, biotic mechanisms also
may be responsible for the demise of inoculants in
100
10-
en
"c
CD
U
^
ID
0_
0.1-
0.01-
0.001-
Figure 85.
T
10
T 1 T
2O 30 40
Time (days)
Survival of fecal coliforms (FC) and fecal
streptococci (FS) in soil during summer and
winter (adapted from Miller, 1979)
70
soil. This is illustrated by the failure of Rhizobium
japonicum cells inoculated in soil (Miller, 1979).
Figure 86A shows that R. japonicum strain 123
survives when inoculated in sterile soil but declines
rapidly in non-sterile soil. The die-back of strain 123
in non-sterile soil was accompanied by the simul-
taneous population increase of a lytic agent,
presumably a bacteriophage (Miller, 1979) (Figure
86B). Other biotic factors such as the production of
microbially produced toxins or antibiotics, predatory
eucaryotes, parasitic procaryotes, lytic enzymes, etc.,
can influence the success of inoculation efforts
(Miller, 1979).
It is likely that the microbial communities existing
in aquifers represent a climax ecological community.
The organisms are found there because they sur-
vived an extensive period of natural selection and
are best able to occupy the available niche. That is,
Table 13. The Survival of Microorganisms Inoculated into Soil
Manured Soil
Manured and Limed Soil
45
Incubation (Days)
100 0
45
*reinoculated
100
Organism
Penicillium sp.
Actinomycetes cellulosae
Bacillus cereus
Pseudomonas fluorescen
Azotobacter chroococcum
Numbers per Gram Dry Soil x 105
24.7
8.4
23.2
142.8
200.
7.7
0.1
57.4
0.
0, 300"
7.1
0.
49.3
0.
0.
33.9
7.6 ''
86.9
.175.
360. •
2.7
0.04
8.6
1.1
120. ,
2.2
0.
12.3
0.
0.
97
-------
10 2O 30 40
Time (days)
Phage
50
_ 0 2 4 6 8 10 12 14
Incubation (days)
Figure 86. (a) The survival of Rhizobium japonicum
strain 123 in sterile and nonsterile soil; (b)
The relationship between the dieback of
strain 123 and the increase in the population
of a bacteriophage (adapted from Miller,
1979).
the indigenous organisms are best able to assume the
functions or "occupations" (Miller, 1979) of organ-
isms in that habitat. Gause's ecological principle
states that only one species can occupy a specific
niche in a habitat (Gause, 1934). Once the available
niche^is filled, climax ecological communities tend to
exhibit the property of homeostasis. The community
tends to be both quantitatively and qualitatively
stable when subjected to moderate levels of biotic or
abiotic stresses.
Little is known about the mechanisms of homeo-
stasis in aquifers. However, it seems certain that
multiple mechanisms, both biotic and abiotic, will
serve to maintain homeostasis. The inoculation of
foreign microorganisms can be viewed as biotic
stress and the mechanisms responsible for the
elimination of these organisms can be perceived as
part of the homeostatic controls. When viewed in this
context, it is unreasonable to expect competitive
success from an inoculant that was likely grown to
high numbers in the laboratory and subsequently
forced to compete in situ with indigenous organisms
naturally selected for their ability to survive the
adverse conditions of nature.
This view on the potential for success of microbial
inoculants in the terrestrial subsurface is pessimistic
and largely based on ecological principles.
Supporting this view though are the problems (not
dealt with here) associated with the transport of
microorganisms in aquifers and the acceptability of
this practice to the regulatory community. Still,
inoculants may occupy a significant role in pollution
mitigation scenarios. For example, homeostatic
control mechanisms of complex environments can be
overwhelmed and frequent inoculation can take
place in order to achieve some desired result. Thus, it
may be possible to superimpose alternate environ-
mental conditions on polluted aquifers with the aim
of selectively favoring a desirable inoculant. How-
ever, inoculants will prove most useful when the
contaminant of interest is exotic and difficult to
degrade by the indigenous microflora. Further, given
the difficulties in transporting bacteria in the
subsurface, inoculants might be most useful in
above-ground and contained treatment processes.
Conclusion
The astonishing metabolic versatility of micro-
organisms has fueled a great deal of excitement
among regulators, researchers, and business people
regarding economical methods for the restoration of
contaminated environments such as aquifers.
However, as with any technology, bioremediation
has both promises and limitations. It is critical to the
future development of this technology that the
practitioners, clients, and regulators recognize the
problems and promises of bioremediation; only with
this recognition can biorestoration be properly con-
sidered as another part of the pollution mitigation
arsenal.
References
Bayly, R. C. and M. G. Barbour, 1984. "The
Degradation of Aromatic Compounds by the Meta
and Gentisate Pathways." In: Microbial Degrada-
tion of Organic Compounds, D. T. Gibson, Editor.
Marcel Dekker, Inc., New York, pp. 253-294.
Brown, J. F. Jr., et al., 1987. "Polychlorinated
Biphenyl Dechlorination in Aquatic Sediments."
Science, Vol. 236, pp. 709-712.
Brown, J. R. Jr., et al., 1984. "PCB Transformations
in Upper Hudson Sediments." Northeast Environ
Sci.,Vol. 3, pp. 167-169.
98
-------
Gause, G. F., 1934. The Struggle for Existence.
Williams and Wilkins, Baltimore, MD.
Gibson, D. T. and V. Subramanian, 1984. "Microbial
Degradation of Aromatic Hydrocarbons." In:
Microbial Degradation of Organic Compounds, D.
T. Gibson, Editor. Marcel Dekker, Inc., New York,
pp. 181-252.
Horowitz, A., J. M. Suflita, and J. M. Tiedje, 1983.
"Reductive Dehalogenations of Halobenzoates by
Anaerobic Lake Sediment Microorganisms." Appl.
Environ. Microbiol., Vol. 45, pp. 1459-1465.
Horvath, R. S., 1972. "Microbial Co-metabolism and
the Degradation of Organic Compounds in
Nature."Bacterial. Rev., Vol. 36, pp. 146-155.
Katznelson, H., 1940a. "Survival of Azotobacter in
Soil." Soil Sci., Vol. 49, pp. 21-35.
Katznelson, H., 1940b. "Survival of Microorganisms
Introduced into Soil." Soil Sci., Vol. 49, pp. 283-
293.
Lee, M. D., et al., 1988. "Biorestoration of Aquifers
Contaminated with Organic Compounds." CRC
Crit. Rev. Environ. Control., Vol. 18, pp. 29-89.
Linkfield, T. G., Suflita, J. M., and J. M. Tiedje, 1989.
"Characterization of the Acclimation Period Prior
to the Anaerobic Biodegradation of Haloaromatic
Compounds." Appl. Environ. Microbiol., (in press).
Miller, R. H., 1979. "Ecological Factors Which Influ-
ence the Success of Microbial Fertilizers or
Activators." Dev. Ind. Microbiol., Vol. 20, pp. 335-
342.
Nelson, M. J. K., et al., 1987. "Biodegradation of
Trichloroethylene and Involvement of an Aromatic
Biodegradative Pathway." Appl. Environ. Micro-
biol., Vol. 53, pp. 949-954.
Nelson, M. J. K., et al., 1986. "Aerobic Metabolism of
Trichloroethylene by a Bacterial Isolate." Appl.
Environ. Microbiol., Vol. 52, pp. 383-384.
Perry, J. J., 1984. "Microbial Metabolism of Cyclic
Alkanes." In: Petroleum Microbiology, R. M. Atlas,
Editor, Macmillan Publishing Co., New York, pp.
61-97.
Quensen, J. F. Ill, J. M. Tiedje, and S. A. Boyd, 1988.
"Reductive Dechlorination of Polychlorinated
Biphenyls by Anaerobic Microorganisms from
Sediments." Science, Vol. 242, pp. 752-754.
Singer, M. E. and W. R. Finnerty, 1984. "Microbial
Metabolism of Straight-chain and Branched
Alkanes." In: Petroleum Microbiology, R. M. Atlas,
Editor. Macmillan Publishing Co., New York, pp.
1-59.
Trudgill, P. W., 1984. "Microbial Degradation of the
Alicyclic Ring: Structural Relationships and Meta-
bolic Pathways." In: Microbiol. Degradation of
Organic Compounds, D. T. Gibson, Editor. Marcel
Dekker, Inc., New York, pp.131-180.
Van Donsel, F. J., E. E. Geldreich, and N. A. Clarke,
1967. "Seasonal Variations in Survival of Indicator
Bacteria in Soil and Their Contribution to Storm
Water Pollution." Appl. Microbiol., Vol. 15, pp.
1362-1370.
Wilson, J. T. and B. H. Wilson, 1985. "Biotrans-
formation of Trichloroethylene in Soil." Appl.
Environ. Microbiol., Vol. 49, pp. 242-243.
You, I.-S. and R. Bartha, 1982. "Stimulation of 3,4-
dichloroaniline Mineralization by Aniline." Appl.
Environ. Microbiol., Vol. 44, pp. 678-681.
99
-------
-------
CHAPTER 9
MODELING SUBSURFACE CONTAMINANT TRANSPORT AND FATE
Joseph F. Keely
Introduction
When scientists and engineers attempt to simulate
the effects of natural phenomena, they are engaging
in modeling. Models are simplified representations of
real-world processes and events, and their creation
and use require many observation-based judgments.
The key theme that drives and focuses the
development and application of models is the faithful
simulation of the specific natural processes. Such
simulation must be demonstrated under a variety of
defined conditions and must incorporate known
scientific facts before a model can be considered
reliable. Many forms of models exist, each having
specific advantages and disadvantages.
Physical models, such as sand-filled tanks used to
simulate aquifers (Figures 87 and 88) and laboratory
columns used to study the relative motion of various
contaminants flowing through aquifer materials
(Figure 89), provide an element of reality that is
enlightening and satisfying from an intuitive
viewpoint. The main disadvantages of physical
models are the extreme effort and time required to
generate a meaningful amount of data. Other
difficulties relate to the care required to obtain
samples of subsurface material for the construction
of these models, without significantly disturbing the
natural condition of the samples.
Analog models also are physically based, but their
operating principle is one of similarity, not true-life
representation. A typical example is the electric
analog model (Figures 90 and 91), where capacitors
and resistors are able to closely replicate the effects
of the rate of water release from storage in aquifers.
As is the case with other physically based models,
Figure 87. Large "sand tank" physical aquifer model.
(The model is constructed of glass walls and
external metal braces. A thick layer of silty
loam overlies a layer of fine sand, which, in
turn, overlies a layer of clay. Scores of
stainless-steel piezometers penetrate the
three layers.)
Figure 88. Close-up view of piezometers in a large
"sand tank" physical aquifer model. (Shallow,
intermediate, and deep piezometers are
wired together in bundles. Wax sheets are
molded over the piezometer bundles to
protect them between samplings.)
101
-------
data generation is slow and little flexibility exists for
experimental design changes.
Figure 89. Laboratory column housed in constant-
temperature environmental chamber. (Con-
taminated solutions are injected into column
through inlet tubing in top, by action of
hydraulic press in foreground. Samples of
the advancing front are withdrawn through
ports visible on right-hand side and bottom
of column.)
Figure 90. Electric analog aquifer model constructed by
Illinois State Water Survey. (The regular array
of resistors and the visible electric "pump"
are hard-wired into a board papered with the
appropriate geologic map.)
Mathematical models are non-physical and rely on
the quantification of relationships between specific
parameters and variables to simulate the effects of
natural processes (Figure 92). Because of this,
mathematical models are abstract and typically
provide little in the way of an intuitive link to real-
world situations. Despite this, mathematical models
can generate powerful insights into the functional
dependencies between causes and effects in the real
world. Large amounts of data can be generated
quickly, and experimental modifications made with
minimal effort, making it possible for many
situations to be studied in great detail for a given
problem.
1. NEAR - SOURCE ZONE
2. INTERMEDIATE ZONE
3, NEAR • W6UF16I.O ZONE
1. NEAR-SOURCE ZONE
Z-tNTEHMEOIATEZOfJE
3. NEAR • WEU-FIELO ZONE
Figure 91. Control panel for electric analog model
shown in Figure 90.
Figure 92. Typical ground-water contamination scenario
and a possible contaminant transport model
grid design for its simulation. (Values for
natural process parameters would be
specified at each node of the grid in
performing simulations. The grid density is
greatest at the source and at potential
location.)
102
-------
Categories of Models
Appropriate models do not exist yet for many
subsurface contamination problems because a
number of natural processes have yet to be fully
understood. This is especially true for transport and
fate evaluations, where chemical and biological
processes are very important but are still poorly
defined. Although great advances are being made in
understanding the behavior of individual contam-
inants, studies of the interactions between contam-
inants are still in their infancy. Also, the current
understanding of physical processes, such as multi-
phase transport, transport through fractured rock,
and transport through karst aquifers, lags behind
needed knowledge. Moreover, certain well-under-
stood phenomena, such as the effects of partially
penetrating wells in unconfined (water-table)
aquifers under varying pumping rates, pose unre-
solved difficulties for mathematical formulations.
A variety of schemes exist for categorizing the
technical underpinnings and capabilities of mathe-
matical models, but the following categorization can
be useful (Bachmat, et al., 1978; van der Heijde, et
al., 1985): -.
• Parameter identification models
• Prediction models
• Resource management models
• Data manipulation codes
Parameter identification models most often are used
to estimate the aquifer coefficients that determine
fluid flow and contaminant transport characteristics,
for example, annual recharge (Puri, 1984), coef-
ficients of permeability and storage (Shelton, 1982;
Khan, 1986a and b), and dispersivity (Guven, et al.,
1984; Strecker and Chu, 1986). Prediction models are
the most numerous kind of model because they are
the primary tools used for testing hypotheses
(Andersen, et al., 1984; Mercer and Faust, 1981;
Krabbenhoft and Anderson, 1986).
Resource management models are combinations of
predictive models, constraining functions (e.g., total
pumpage allowed), and optimization routines for
objective functions (e.g., scheduling wellfield opera-
tions for minimum cost or minimum drawdown/
pumping lift). Very few of these models are devel-
oped and supported enough to be considered practi-
cally useful and there does not appear to be a
significant drive to improve this situation (van der
Heijde, 1984a and b; van der Heijde, et al., 1985).
Data manipulation codes also received little
attention until only recently. These codes are now
becoming increasingly popular because they simplify
data entry (e.g., preprocessors) to other kinds of
models and facilitate the production of graphic dis-
plays (e.g., postprocessors) of model outputs (van der
Heijde and Srinivasan, 1983; Srinivasan, 1984;
Moses and Herman, 1986). Other software packages
are available for routine and advanced statistics,
specialized graphics, and database management
needs (Brown, 1986).
Quality Control
Quality control measures are greatly needed for
modeling the transport and fate of subsurface con-
taminants, particularly in the use of numerical
models. Huyakorn, et al. (1984) suggested three
levels of quality control:
1. Validation of the mathematical basis of a model
by comparing its output with known analytical
solutions to specific problems.
2. Verification of the applicability of a model to
various problem categories by successful simu-
lation of observed field data.
3. Benchmarking the efficiency of a model in
solving problems by comparison with the per-
formance of other models.
These levels of quality control address the soundness
and utility of the model alone but do not treat
questions of its application to a specific problem.
Hence, at least three additional levels of quality
control appear justified:
4. Critical review of the problem conceptualization
to ensure that the modeling effort considers all
physical, chemical, and biological processes that
may affect the problem.
5. Evaluation of the specifics of the model's applica-
tion, e.g., appropriateness of the boundary condi-
tions, grid design, time steps, etc.
6. Appraisal of the match between the mathe-
matical sophistication of the model and the
temporal and spatial resolution of the data.
Validation of the mathematical framework of a
numerical model is deceptively simple. The usual
approach for ground-water flow models involves a
comparison of drawdowns predicted by the Theis
analytical solution to those obtained by using the
model. The deceptive part is that the Theis solution
can treat only simplified situations as compared with
the scope of situations addressable by the numerical
model. In other words, analytical solutions cannot
test most of the capabilities of numerical models in a
meaningful way; this is particularly true in simula-
ting complex aquifer boundaries and irregular
chemical distributions.
103
-------
Field verification of a numerical model consists of
two steps, calibrating the model using one set of
historical records (e.g., pumping rates and water
levels from a certain year), and then, attempting to
predict a subsequent set of observations. In the
calibration phase, the aquifer coefficients and other
model parameters are adjusted to achieve the best
match between model outputs and known data. In
the predictive phase, no adjustments are made
except for actual changes in pumping rates, etc.
Presuming that the aquifer coefficients and other
parameters are known with sufficient accuracy, a
mismatch means that the model either is not
correctly formulated or does not treat all of the
important phenomena affecting the actual field
situation, such as leakage between two aquifers.
Field verification usually leads to additional data-
gathering efforts because existing data for the
calibration procedure often are insufficient to pro-
vide unique estimates of key parameters. This means
that a black box solution may be obtained, which is
valid only for the observation period used in the
calibration. For this reason, the blind prediction
phase is an essential check on the uniqueness of the
parameter values used in the model. In this regard,
field verification of models using datasets from
controlled research experiments may be more
practical to achieve than with the data generated
during a Superfund site investigation.
Benchmarking routines are available that compare
the efficiency of different models in solving the same
problem (Ross, et al., 1982; Huyakorn, et al., 1984);
however, more must be done in this area. For
example, common observations indicated that finite
element models (FEMs) have an inherent advantage
over finite difference models (FDMs) in terms of
ability to incorporate irregular boundaries (Mercer
and Faust, 1981) (the number of points (nodes) used
by FEMs is considerably less due to the flexible nodal
spacings allowed). Benchmarking routines, however,
show that the large amount of computer time
required to evaluate FEM nodes reduces the cost
advantage for simulations of comparable accuracy.
A Field Sample
Field experience using special geotechnical methods
and state-of-the-art research findings was gained at
the 20-acre Chem-Dyne solvent reprocessing site in
Hamilton, Ohio (Figure 93), where over 250 chemical
waste generators disposed of drummed or bulk
wastes during its operational lifetime (1974-1980).
Poor waste handling practices, such as purposeful
pn-site spillage of a wide variety of industrial chem-
icals and solvents, direct discharge of liquid wastes
to a stormwater drain beneath the site, and mixing of
incompatible wastes, occurred routinely at Chem-
Dyne. These practices caused extensive soil and
ground-water contamination, massive fish kills in
the Great Miami River, and major on-site fires and
explosions.
The stockpiling of liquid and solid wastes resulted in
a long-term threat to the environment. More than
50,000 drums of hazardous waste were stored at the
site at its peak of operations (CH2M-Hill, 1984a).
The drums were stacked improperly, in tiers five and
six drums high, causing the drums at the bottom to
buckle and corrode. After the remedial investigation
began in the spring of 1982, more than 20,000 drums
still remained; at least 8,500 of these were so badly
corroded that they could not be identified. A number
of bulk chemical storage tanks also were abandoned
on site. Visual observations indicated that raw
chemical salts and oils had been poured out on the
sand-and-gravel ground surface.
The FIT Investigation
The seriousness of the ground-water contamination
problem at Chem-Dyne became evident during the
initial site survey (1980-1981), which included the
construction and sampling of over twenty shallow
monitoring wells (Ecology and Environment, 1982).
The initial survey indicated that the contaminant
problem was much more limited than was later
shown to be the case (Roy F. Weston Inc., 1983;
CH2M-Hill, 1984a). A good portion of the improve-
ment in delineating the plume was brought about by
a better understanding of the natural processes
controlling transport of contaminants at the site.
The initial site survey indicated that ground water
flowed to the west of the site (toward the Great
Miami River), but that a shallow trough paralleled
the river as a result of weak and temporary stream
influences. The study concluded that contaminants
already in the aquifer would be discharged into the
river and would not need to be removed (Ecology and
Environment, 1982). That study also concluded that
the source was limited to highly contaminated
surface soils, and that removal of the uppermost
three feet of the soil would essentially eliminate the
source of contaminants.
That conclusion, however, was based on faulty soil
sampling procedures. The soil samples taken were
not preserved in air-tight containers, so most of the
volatile organic chemicals leaked out prior to
analysis. The uppermost soil samples probably
showed high volatile organic levels because of the co-
occurrence of viscous oils and other organic chem-
icals that may have served to entrap the volatiles.
The more viscous and highly retarded chemicals did
not migrate far enough into the vertical profile to
104
-------
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
Location Map
LEGEND
• monitoring well locations
MW1 monitoring well Identity
• • • site boundary
Figure 93. Monitoring well location map for the Chem-Dyne site.
exert a similar influence on samples collected at
depths greater than a few feet.
Additional Site Investigations
Subsequent studies of the site corrected these
misinterpretations by producing data from proper
soil samplings and incorporating much more detailed
characterizations of the fluvral sediments and the
natural flow system. In those studies, vertical profile
characterizations were obtained from each new
borehole drilled by continuous split-spoon samples of
subsurface solids. The split-spoon samples helped to
confirm the general locations of interfingered clay
lenses and clearly showed the high degree of
heterogeneity of the sediments (Figures 94 and 95).
For example, a dense clay lens was found at similar
elevations (at 570 to 580 ft MSL) along the valley
axis (Figure 94), but was found only intermittently
perpendicular to the valley axis (Figure 95). This
should be expected by hydrogeologic intuition since,
during flooding, rivers dissect deposits that were laid
down in low energy periods and rivers undergo
natural channel changes (e.g., meandering) as they
mature. These phenomena create lenticular clay
bodies of very limited extent and structurally
anisotropic conditions. The major axis of anisotropy
parallels the average downstream direction, which
itself is generally co-linear with the valley axis.
Clusters of vertically separated monitoring wells
were constructed during the remedial investiga-
tions/feasibility studies at Chem-Dyne. While an
extensive network of shallow wells confirmed earlier
indications of general ground-water flow toward
Great Miami River (Figure 96), the clusters of
vertically separated wells revealed that downward
gradients existed adjacent to the river. Figures 97
and 98 show that these vertical gradients, which
105
-------
CHEM-DYNE GEOLOGIC CROSS-SECTION
NNW
V)
o
600
590
580
570
560
550
540
530
SSE
-J 520
111
510
500
m
Water Level Elevation" ]
Approximately 563 ft. MSL j
-October 30, 1983
DATA SOURCE: CH2M-HIII, 1984a
Fill (sandy gravel)
Clayey silt, silty clay
Sand
Sandy gravel, gr. sand
Silty sand
f-Xi Clayey gravel, glacial till
Figure 94. NNW-SSE geologic cross-section at Chem-Dyne site.
ranged from a 1- to 3-foot drop over the 20-foot
vertical separation between the bottom of the shal-
low wells and the top of the screens in the deep wells
(or about 0.100), are quite dramatic relative to the
horizontal gradient across the contaminant plume
(which averages about 0.001). This finding indicated
that the migrating plume would not be discharged to
the river, but would flow under the river.
The presence of major industrial wells on the west
bank of Great Miami River provided an explanation
for the observed downward vertical gradients
(normally, one would have expected the river to be
gaining water from the aquifer at this point in the
basin), and supported the conclusion that
contaminants could not be discharged to the river
from the aquifer. The plume would be drawn to
greater depths in the aquifer by the locally severe
downward gradient, but it could not be determined if
the industrial wells would actually capture the
plume. That determination would require careful
evaluation of the hydrogeologic features beneath the
river (an activity not attempted because of the
associated costs) and expectations that planned
remedial actions would stop the plume before
substantial encroachment could occur.
Hydrologic Complications
Unfortunately, the hydraulic interplay of the river
with the aquifer was not well appreciated by the field
crews taking routine measurements of water levels.
Observations late in the final study found that,
during preparations for a pump test, river stage
variations cause as much as three feet of water level
change during a single day at wells close to the river.
This effect was virtually negligible at wells much
closer to the site. This sort of situation makes it
crucial to obtain water levels at all wells within only
a few hours, otherwise, the sort of confusing water
level maps shown in Figures 99 to 101 may result.
These figures were prepared with water-level
elevation data that were measured over periods of
several days (CH2M-Hill, 1984a).
Investigators decided to use a major pump test to
estimate the hydrogeologic characteristics of the
106
-------
600
590
O
_J 580
21 570
560
CHEM-DYNE GEOLOGIC CROSS-SECTION
WSW ENE
A
O
550
540
53D
&
>: Clayey gravel, glacial till
Figure 95. WSW-ENE geologic cross-section at Chem-Dyne site.
heavily contaminated portion of the aquifer. The
pump test was technically difficult because the
pumping well had to be drilled on site due to
potential liability and lack of access elsewhere. The
'drillers were substantially slowed by the need to
wear air-tanks when encountering particularly con-
taminated subsoils that emitted volatile fumes and
presented unacceptable health risks. Since the
pumped water was expected to.be contaminated, ten
large contemporary holding tanks (100,000 gallons
each) were constructed on site to impound the waters
for testing and possible treatment before being
discharged to the local sewer system (CHgM-Hill,
1984a).
Although there had been some resistance to
conducting the pump test because of its cost, the test
results were very valuable. The water levels in
thirty-six monitoring wells were observed during the
test, providing a very detailed picture of area! trans-
missivity variations (Figures 102). This information
helped explain the unusual configuration of the
plume shown in Figures 103, 104, and 106 (Figure
105 is an updated location map for 1985 data
presented in Figure 106 and some later figures). The
information also was used to guide the design of a
pump-and-treat system. Storage coefficients also
were estimated and, though the short duration of the
test (14 hours) did not provide many definitive
estimates, qualitative confirmation of the generally
non-artesian (water-table) nature of the aquifer
beneath the site was clearly confirmed, as were the
increasingly artesian conditions from the west edge
of the site towards the river.
Anisotropic Flow Biases
On-site transmissivity estimates from a trio of wells
(MW-23, MW-26, and MW-29) indicated a 2:1 aniso-
tropic bias toward the river as opposed to
downvalley, whereas nearer the river a second trio of
wells (MW-28, MW-33, and MW-35) yielded esti-
mates for which the bias appeared to be 10:1
downvalley (CH2M-Hill, 1984a). These trends coin-
cide with the nature of the system; that is, there are
few clay occurrences on site and east of the site from
whence recharge waters flow toward the river, and
107
-------
HYDRAULIC CANAL
I
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
April, 1983 Water Level Elevation Contours
for Shallow Wells
(values in feet)
Figure 96. April 1983 water level elevation contours at Chem-Dyne site using shallow well data only, in feet MSL.
significant occurrences of clays at the west fenceline
of the site and adjacent to the river. The latter clays
had lenticular shapes that paralleled the axis of the
river, causing a strong downvalley bias. These
indications mean that flow would not proceed at an
average velocity perpendicular to the water level
contours that would be west from the northern half of
the site, and south to southwest from the southern
half of the site (Figure 95). Instead, flow would move
westerly first and then southerly as it neared the
river. By the time areas adjacent to the river were
reached, the net position would be roughly the same,
but the path taken to get there would be strikingly
different. The contaminant masses leaving both the
northern and southern halves of the site would be
concentrated in west-trending tongues. Moreover,
the future travel paths with the anisotropic bias
interpretation would not be the same as with a
presumption of isotropicity.
The distributions of contaminants observed at Chem-
Dyne seem to support this anisotropic bias inter-
pretation. The highest concentrations lie along an
axis that does not appear to be influenced by the
southerly components of flow expressed by the water-
level contours offsite (Figures 103 and 104). While it
is true that the pumpage of major production wells
on the other side of the river offers a nominal
explanation for this, the actual water level contours
contradict the notion that the industrial wells
dominate the entire flow field. This has to be
appreciated in terms of capture zones; the industrial
wells will indeed affect all of the ground-water flow-
lines locally, but may not capture all of them. It is
likely that the plume is bifurcating near the river,
with one member travelling on flowlines captured by
the industrial wells, and a second member tangen-
tially affected, but eventually released to continue on
down the valley with the rest of the aquifer waters.
108
-------
570
560
550
< 540
111
530
520
MB-20
MIU-19
D
DEO 82 APR 83 JUN 83 OCT 83
HELL SCREENS
COMPARISON OF WATER LEVEL ELEVATIONS IN
MONITORING WELLS MW-19 AND MW-2D
figure 97. Water levels in a cluster of two vertically
separated monitoring wells adjacent to Great
Miami River and due west of Chem-Dyne site.
570
560
550
o -.. . ,
'< 540
ILI '
530
520
rw-21
MU-22
D
DEC 82 APR 83 JUN 83 OCT 83 HELL SCREENS
. ( COMPARISON OF WATER LEVEL ELEVATIONS IN
MONITORING WELLS MW-21 AND MW-22
Figure 98. Water levels in a cluster of two vertically
• ..- . • separated monitoring wells at the confluence
•'.•' of Great Miami River and Ford Hydraulic
Canal, due west of north boundary of Chem-
Dyne site.
Field Evidence for Biotransformations
Finally, the distribution patterns of contaminant
species that emerged from the investigations at
Chem-Dyne were understood by considering re-
search results and theories regarding chemical and
microbiological influences. Contaminant distribu-
tion maps derived from samples taken at the end of
the field investigation (October, 1983, only months
after the last drums of solvents were removed from
the site) suggested that the transformation of
tetrachloroethene (Figure 107) to less halogenated
daughter products such as trichloroethene (Figure
108), dichloroethene (Figure 109), and vinyl
chloride/monochloroethene (Figure 110) was occur-
ring.
In such circumstances, one would expect to see the
progressive disappearance of tetrachloroethene and
successive increases in the concentrations and extent
of potential daughter products. This seems to be the
case at Chem-Dyne, according to the October 1983
data. One might argue that too little vinyl chloride is
observed (Figure 110) to show the full series of
degradation expected, but there are plausible
reasons why the distributions might be as shown.
For example, with a continuous source input, the.
concentrations of tetrachloroethene might be high
enough that there would be no need for further
biotransformation of daughter products because an
ample food supply is available in the parent mate-
rial. Alternatively, the concentrations of tetrachloro-
ethene with continuous source inputs might indeed
be so high as to limit biotransformation by toxic
effects. Since the relative kinetics of the various
transformations in this sequence still require further
definition, it is impossible to make rigorous conclu-
sions with regard to these possibilities.
But consider the data obtained .during a chemical
sampling conducted two years later in preparation
for activation of the pump-and-treat system used to
remediate the plume (Figures 110 to 113). At least
two years of freedom from surface inputs of solvents
had occurred as well as two years of healthy rain-
water flushing the unsaturated zone of stored
residues. Investigators found that the daughter
products (Figures 111 to 113) contain much greater
mass than that in the tetrachloroethene contours
(Figure 110), and are spread over significantly
greater areas. The increase in the vinyl chloride
component of the plume is staggering (compare
Figure 113 to Figure 109). These data highly suggest
active degradation of tetrachloroethene to its
possible daughter products. Knowledge of this kind
of possible transformation should be valuable to
those attempting to design and estimate costs for
treatment systems since treatment efficiencies vary
with the contaminant and its contribution to the
109
-------
HYDRAULIC CANAL
I
LEGEND
• monitoring well location
MW1 monitoring well Identity
• • • site boundary
x water level elevation contour
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
December, 1982 Water Level Elevation Contours
for Shallow Wells
(values in feet)
Figura 99. December 1982 water level elevation contours at
overall loading. Vinyl chloride is much more volatile
and easier to remove than tetrachloroethene.
The relative rates of movement of these and other
common solvents like benzene and chloroform at
Chem-Dyne generally conformed to predictions
based on sorption principles. This is useful in a
practical sense. The remediation efforts made use of
these contaminant transport theories in estimating
the capacity of the treatment system and the length
of time needed to remove residuals from the aquifer
solids (CH2M-Hill, 1984b).
The Role of Mathematical Models
During the latter stages of negotiations with the
Potentially Responsible Parties (PRPs), the State of
Ohio government contractors prepared mathe-
matical models of the flow system and contaminant
Chem-Dyne site using shallow well data only in feet MSL.
transport at Chem-Dyne (GeoTrans, 1984). These
models were used to estimate the possible direction
and rate of migration of the plume in the absence of
remediation, the mass of contaminants removed
during various remedial options, and the effects of
sorption and dispersion on those estimates. Because
of the wide range of sorption properties associated
with the variety of contaminants found in significant
concentrations, it was necessary to select values of
retardation constants that represented the likely
upper and lower limits of sorptive effects. It also was
necessary to estimate or assume the values of other
parameters such as dispersion coefficients known to
affect transport processes.
Ward and his co-workers recently published this
application of models to the Chem-Dyne site in an
article highlighting the development and recom-
mendation of the Telescopic Mesh Refinement (TMR)
110
-------
LEGEND
• monitoring well location
MW1 monitoring well identity
• • • site boundary
«»••»->• water level elevation contour
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
June, 1983 Water Level Elevation Contours
for Shallow Wells
(values in feet)
Figure 100. June 1983 water level elevation contours at ChenvOyne site using shallow well data only, in feet MSL.
modeling approach (Ward, et al., 1987). The TMR
approach involves a staged evaluation of the prob-
lem, proceeding from the regional level to the local
level to the site-specific level (Figure 114). This
approach makes it possible to assure the appro-
priateness and consistency of boundary conditions
such as recharge rates to the aquifer and interactions
with streams. GeoTrans chose to develop this set of
models using the SWIFT model, which utilizes the
finite difference method (FDM), that, in turn, utilizes
rectangular grids (Figure 115). FDM is the most
mathematically straightforward and easily de-
bugged of the numerical analysis techniques used for
mathematical modeling (Mercer and Faust, 1981).
Recognizing the need to account for the influences of
the vertical gradients previously noted to be severe
adjacent to Great Miami River, GeoTrans created a
quasi-three-dimensional (layered) model at the local
scale (Figure 116).
The results of their modeling efforts included maps of
the potentiometric surfaces represented by water
level elevations in wells tapping the different model
"layers" (Figures 117 and 118). These maps indicated
that local industrial wells control flow in the deeper
portions of the aquifer. This site-scale model was
used to investigate the probable effectiveness of the
pump-and-treat remediation proposed by the
Potentially Responsible Parties for Chem-Dyne
(Figures 119 and 120). Geo-Trans concluded that the
proposed pumping scheme appeared to be quite
effective in the interior zone of the contaminant
plume, but could not be completely effective in
developing an inward hydraulic gradient at the
111
-------
HYDRAULIC CANAL
LEGEND
• monitoring well location
MW1 monitoring well identity
• • • site boundary
s*-^s water level elevation contour
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
October, 1983 Water Level Elevation Contours
for Shallow Wells
(values in feet)
Figure 101. October 1983 water level elevation contours at Chem-Dyne site using shallow well, data only, in feet MSL.
periphery of the plume. Their predictions turned out
to be correct; during the first year of operation, the
net withdrawal rate of the remediation wellfield was
increased substantially over the originally proposed
values, and inward hydraulic gradient control was
still not fully established (Figure 121). GeoTrans
modeling efforts also included predictions of the
extraction well concentrations versus time (Figure
122), but comparisons with actual performance are
difficult because of the many remediation shut-down
periods (including lost time due to clogging of the air-
stripper by precipitated iron and manganese).
Large uncertainties were associated with those
modeling efforts due to lack of information about the
actual history of chemical inputs and other impor-
tant data. However, there was agreement between
the government and PRP technical experts about the
helpfulness of modeling efforts in assessing the
magnitude of the problem and in determining mini-
mal requirements for remediation. Consequently,
modeling efforts continue at Chem-Dyne. Data gen-
erated during the remediation phase are being used
to refine models in an ongoing process so that the
effectiveness of the remedial action can be evaluated
properly. .
Summary
Models of the transport and fate of contaminants in
the subsurface environment are created by organ-
izing known information and relationships into a
functional representation (e.g., a sand-filled tank, a
circuit board, or an equation). Models may be used to
simulate the response of specific problems to a wide
variety of possible solutions.
112
-------
LEGEND
• monitoring well locations
MW1 monitoring well identity
... site boundary
^ux Transmissivity Isopleth in thousands
of square feet per day
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
Transmissivity Estimate from October
1983 Pump Test
(Deep Well Contours Only)
Figure 102! Transmissivity estimates obtained from October 1983 Pump test at Chem-Dyne site, in thousands of
square feet/day (plotted values of estimates from all wells).
The application of mathematical models is subject to
considerable error in practical situations when
appropriate field determinations of natural process
parameters are lacking. Contrary to popular beliefs,
this source of error is not addressed adequately by
sensitivity analyses or by the application of
stochastic techniques for estimating uncertainty.
Rather, the high degree of hydrogeological, chemical,
and microbiological complexity typically present in
field situations forces the use of site-specific charac-
terization of the influences of various natural
processes by detailed field and laboratory investi-
gations.
Both the mathematics that describe models and the
parameter inputs to those models must be subjected
to rigorous quality control procedures. Otherwise,
results from held applications of models are likely to
be qualitatively, as well as quantitatively, incorrect.
Quality control methodologies must focus on the
accuracy of the problem conceptualization and the
representativeness of parameter values, and recog-
nize that accuracy and precision determinations are
insufficient measures of quality.
References
Andersen, P. F., C. R. Faust, and J. W. Mercer, 1984.
"Analysis of Conceptual Designs for Remedial
Measures at Lipari Landfill." Groundwater, Vol.
22, No. 2.
Bachmat, Y., et al., 1978. "Utilization of Numerical
Groundwater Models for Water Resource Manage-
ment." EPA-600/8-78-012, R. S. Kerr Environ-
mental Research Laboratory, U.S. Environmental
Protection Agency, Ada, OK.
113
-------
HYDRAULIC CANAL
if
1
Scali
3OOfl
lOOro
LEGEND
• monjtorjng well locations
MW1 monitoring well Identity
. . . site boundary
Isopleth In parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TOTAL VOLATILE ORGANIC CHEMICALS
APRIL 1983 SAMPLING
Figure 103. April 1983 total VOC concentration contours (ppb) at Chem-Dyne site, using shallow well data only.
Brown, J., 1986. "Environmental Software Review."
Pollution Engineering, Vol. 18, No. 1.
CH2M Hill, 1984a. "Remedial Investigation Report:
Chem-Dyne Site, Hamilton, Ohio." Unpublished
report, Contract No. 68-01-6692 to U.S. EPA
Region 5, Chicago, IL.
CH2M Hill, 1984b. "Feasibility Study Report: Chem-
Dyne Site, Hamilton, Ohio." Unpublished report,
Contract No. 68-01-6692 to U.S. EPA Region 5,
Chicago, IL.
Ecology and Environment, 1982. "Field Investigation
of Uncontrolled Waste Sites — Ground Water
Investigation of Chem-Dyne Sites." Unpublished
report, Contract No. 68-01-6506 to U.S. EPA
Region 5, Chicago, IL.
Faust, C. R., L. R. Silka, and J. W. Mercer, 1981.
"Computer Modeling and Ground-water Protec-
tion." Groundwater, Vol. 19, No. 4.
GeoTrans, 1984. "Evaluation and Analysis of
Groundwater and Soil Contamination at the
Chem-Dyne Site, Hamilton, Ohio." Report to the
Ohio Environmental Protection Agency, October
30,1984.
Guven, O., F. J. Molz, and J. G. Melville, 1984. "An
Analysis of Dispersion in a Stratified Aquifer."
Water Resources Research, Vol. 20, pp. 1337-1354.
Huyakorn, P. S., et al., 1984. "Testing and Valida-
tion of Models for Simulating Solute Transport in
Ground Water: Development and Testing of Bench-
mark Techniques." IGWMC Report No. GWMI84-
114
-------
LEGEND
• monitoring well locations
MW1 monitoring well identity
- - - site boundary
Isopleth in parts per billion
(shallow welfs only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TOTAL VOLATILE ORGANIC CHEMICALS
JUNE 1983 SAMPLING
Figure 104. June 1983 total VOC concentration contours (ppb) at Chem-Dyne site, using shallow well data only.
13, International Ground Water Modeling Center,
Holcolm Research Institute.
Khan, I. A., 1986a, "Inverse Problem in Ground
Water: Model Development." Groundwater, Vol.
24, No. 1.
Khan, I. A. 1986b, "Inverse Problem in Ground
Water: Model Application." Groundwater, Vol. 24,
No. 1.
Krabbenhoft, D. P. and M. P. Anderson, 1986. "Use
of a Numerical Ground-water Flow Model for
Hypothesis Testing." Groundwater, Vol. 24, No.l.
Mercer, J. W. and C. R. Faust, 1981. Ground-Water
Modeling. National Water Well Association,
Worthington, OH.
Moses, C. O. and J. S. Herman, 1986. "Computer
Notes — WATIN — A Computer Program for Gen-
erating Input Files for WATEQF." Groundwater,
Vol. 24, No. 1.
Puri, S., 1984. "Aquifer Studies Using Flow Simula-
tions." Groundwater, Vol. 22, No. 5.
Ross, B., et al., 1982. "Benchmark Problems for
Repository Siting Models." U.S. NRC Publication
No. NUREG/CP-3097, U:S. Nuclear Regulatory
Commission, Washington, DC.
Roy F. Weston, Inc., 1983. "Preliminary Hydro-
geologic Investigation and Preliminary Evaluation
of Remedial Action Alternatives Feasibility:
Chem-Dyne Hazardous Materials Recycling Facil-
ity, Hamilton, Ohio." Unpublished Report, Con-
tract No. 68-03-1613 to U.S. EPA Region 5,
Chicago, IL.
Shelton, M. L., 1982. "Ground-water Management in
Basalts." Groundwater, Vol. 20, No. 1.
115
-------
LEGEND
• monitoring well location
MW1 monitoring well identity
• • - site boundary
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
UP DATED LOCATION MAP
(G-WELLS ADDED)
Figure 105. Updated location map for the Chem-Dyne site.
Srinivasan, P., 1984. "PIG - A Graphic Interactive
Preprocessor for Ground-water Models." IGWMC
Report No. GWMI 84-15. International Ground
Water Modeling Center, Holcolm Research
Institute.
Strecker, E. W. and W. Chu, 1986. "Parameter
Identification of a Ground-water Contaminant
Transport Model." Groundwater, Vol. 24, No. 1.
van der Heyde, P. K. M., 1984a. "Availability and
Applicability of Numerical Models for Ground
Water Resources Management." IGWMC Report
No. GWMI 84-19. International Ground Water
Modeling Center, Holcolm Research Institute.
van der Heyde, P.K.M., 1984b. "Utilization of Models
as Analytic Tools for Groundwater Management."
IGWMC Report No. GWMI 84-18. International
Ground Water Modeling Center, Holcolm Research
Institute.
van der Heijde, P.K.M., et al., 1985. "Groundwater
Management: The Use of Numerical Models, 2nd
Edition." AGU Water Resources Monograph No. 5.
American Geophysical Union, Washington, DC.
van der Heijde, P. K. M. and P. Srinivasan, 1983.
"Aspects of the Use of Graphic Techniques in
Ground-water Modeling." IGWMC Report No.
GWMI 83-11. International Ground Water Model-
ing Center, Holcolm Research Institute.
Wagner, B. J. and S. M. Gorelick, 1987. "Optimal
Groundwater Quality Management Under Param-
eter Uncertainty." Water Resources Research, Vol.
23, No. 7.
Ward, D. S., et al., 1987. "Evaluation of a Ground-
water Corrective Action at the Chem-Dyne Haz-
ardous Waste Site Using a Telescopic Mesh
Refinement Modeling Approach. "Water Resources
Research, Vol. 23, No. 4.
116
-------
HYDRAULIC CAN.AL
LEGEND
- • monitoring well locations
MW1 monitoring well identity
... site boundary
O^s^ Isopleth in parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TOTAL VOLATILE ORGANIC CHEMICALS
DECEMBER 1985 SAMPLING
Figure 106. December 1985 total VOC concentration contours (ppb) at Chem-Dyne site, using shallow well data only.
117
-------
HYDRAULIC CANAL
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TETRACHLOROETHENE
OCTOBER 1983 SAMPLING
LEGEND
• monitoring well locations
MW1 monitoring well identity
site boundary
Isopleth in parts per billion
(shallow wells only)
Figure 107. October 1983 tetrachloroethane concentration contours (ppb) at Chem-Dyne site, using shallow well data
only.
118
-------
HYDRAULIC CANAL
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
LEGEND
.... • monjtormg well locations
MW1 monitoring well identity
- - . site boundary
Isopleth in parts per billion
(shallow wells only)
TRICHLOROETHENE
OCTOBER 1883 SAMPLING
Figure 108. October 1983 trichloroethane concentration contours (ppb) at Chem-Dyne site, using shallow well data
only.
119
-------
LEGEND
• monitoring well locations
MW1 monitoring well identity
... site boundary
^ Isopletn In parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
trans-DICHLOROETHENE
OCTOBER 1983 SAMPLING'"
Figure 109. October 1983 trans-dichloroethene concentration contours (ppb) at Chem-Dyne site, using shallow well
data only.
120
-------
HYDRAULIC CANAL
LEGEND
• monitoring well locations
MW1 monitoring well identity
. . . site boundary
-^*» Isopleth in parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
VINYL CHLORIDE
(MONOCHLOROETHENE)
OCTOBER 1983 SAMPLING
Figure 110. October 1983 vinyt chloride concentration contours (ppb) at Chem-Dyne site, using shallow well data only.
121
-------
LEGEND
• monitoring well locations
MW1 monitoring well identity
... site boundary
Isopleth in parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TETRACH LOROETH EN E
DECEMBER 1985 SAMPLING
Figure 111. December 1985 tetrachloroethene concentration contours (ppb) at Chem-Dyne site, using shallow well
data only.
122
-------
I
1
LEGEND
..,.,* monitoring well locations
MW1 monitoring well Identity
... site bounda-
--N^ Isopletl '
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
TRICHLOROETHENE
DECEMBER 1985 SAMPLING
i boundary
pleth In parts per billion
(shallow wells only)
w Figure 112.
December 1985 trichloroethene concentration contours (ppb) at Chem-Dyne site, using shallow well data
only.
123
-------
HYDRAULIC CANAL
LEGEND
• monitoring well locations
MW1 monitoring well identity
• - . site boundary
Isopleth in parts per billion
(shallow wells only)
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
trans-1-2-DICHLOROETHENE
DECEMBER 1985 SAMPLING
Figure 113. December 1985 trans-dichloroethene concentration contours (ppb) at Chem-Dyne site, using shallow well
data only. ,
124
-------
HYDRAULIC CANAL
CHEM-DYNE SUPERFUND SITE
Hamilton, Ohio
VINYL CHLORIDE
(MONOCHLOROETHENE)
DECEMBER 1985 SAMPLING
LEGEND
.... * monitoring well locations
MW1 monitoring well Identity
. . . site boundary
P -••*«"• soofi.
Isopleth in parts per billion
(shallow well
Figure 114. ,.December. 1985 yinyl chloride concentration contours (ppb) at Chem-Dyne site, using shallow well data
only. • ..".-.'
125
-------
REGIONAL
200 FEET
0 30 CO UETEOS
Figure 115. Conceptual diagram of the telescopic mesh refinement modeling approach (Ward, et al., 1987).
126
-------
Figure 116.
Finite-difference grid used for the regional-
scale flow model of the Great Miami River
Valley-fill aquifer (Ward, et al., 1987).
STREAM
REACHES
HAMILTON
NORTH WELLFIELD
B CHAMPION
PAPER
LOCAL MODEL
CONSTANT PRESSURE
BOUNDARY
• PUMPING CENTER
HAMILTON
SOUTH WELLFIELD
127
-------
Figure 117. Local-scale model conceptualization
(not to scale) (Ward, et al., 1987).
TWO-MILE
0AM
GREAT
MIAMI
RIVER
FORO
HYDRAULIC
CANAL
CHAMPION
PAPERS
MODEL
LAYER
J_
~3
4
5
•Kh^ooft/d:
jijirty ;Sands';and 'firbvelV:-::;
SiUs'Sands',Clays'' '• '-'•••.' '' ',".
•. . • . . . •. c!ea.n Sand and Gravels.
rNOTE
: Conversion factor
1ftd~':0.3049md'}
128
-------
i r-j ri-7 :;--:•'" :-•••••- •.
_M ! i i i '•$..;! | i i! j <
till i iff hrr ri"1
577
HYDROELECTRIC DAM—
561
562
564
563
/CALCULATED
«V POTENTIOMETRIC
<§* SURFACE CONTOUR
/ IN FEET
• PUMPING CENTER
o 1000 FEET
HV-H
0 300 METERS
NOTE: Conversion Factor
Iff* 0.3049m
Figure 118. Calculated potentiometric surface for the
shallow interval (local-scale flow model,
layer 1) (Ward, et al., 1987).
129
-------
563
561
562
/CALCULATED
A/ POTENTIOMETRIC
«T SURFACE CONTOUR
/ IN FEET
• PUMPING CENTER
0 1OOO FEET
I I ' I I
0 300 METERS
NOTE: Conversion Factor
1ft= 0.3049m
si
•
•X
4:
|
i
1
• i
I
i
l
1 0
1
. REMEDIATION WELLS
. R
- s
E:
ii
. E
II
»TE IN 6PM IS
HALLOW
(TRACTION 0
4TERMEDIATE
(TRACTION A
SECTION •
1
1
~
;
:
\
j'
f
•
-
20
\m
\
to
!jl
u
i
o
It
49
-,14
a
it
1 MOTS •' Ccnnnmn Fxur
l«w> • OMJILfi
|
*
bl
L
Bis
•^
:
•*
o2
u
V
V
i
0 100 «0 FttT
o 30 (o Hrrcn
Figure 120. Finite-difference grid used in the site-scale
flow and transport model for the Chem-
Dyne site. Relationship to local-scale
model is shown in Figure 115 (Ward, et al.,
1987).
Figure 119. Calculated potentiometric surface for the
deep interval (local-scale flow model, layer
4) (Ward, etal., 1987).
130
-------
MASS UNREMEDIATED
LEAVING GRID
PROPOSED
PLAN
\ \MASSREMOVED
^» X DV OCIiCniATI.
BY REMEDIATION
MASS IN GROUNDWATER
H0it'- Conversion Factor
lib - 0.4535
100
TIME (doys)
Figure 121. Nviass of VOCs versus time (Ward, et al., 1987).
1000
10,000
0.1 1 10
TIME (days)
1000
Figure 122. Extraction well concentrations versus time
(Ward, etal., 1987).
131
-------
-------
CHAPTER 10
MANAGEMENT CONSIDERATIONS IN TRANSPORT AND FATE ISSUES
Joseph F. Keely
Perspectives for Site Characterizations
and Remediations
The preceding chapters discussed many concepts
pertinent to investigating and predicting the trans-
port and fate of contaminants in the subsurface.
Recognition of the fact that these concepts are
evolving is important when making decisions
regarding detection and remediation of subsurface
contamination. This importance lies not only in an
awareness of existing uncertainties, but also in the
realization that conventional site characterization
approaches have fallen considerably behind the
state-of-the-art. From a practitioner's perspective,
most subsurface contamination assessments do not
adequately emphasize the need to obtain detailed
information about preferential pathways and the
natural processes affecting the transport behavior
and ultimate fate of contaminants. When site
characterization efforts incorporate state-of-the-art
characterizations of natural process parameters,
rather than relying almost exclusively on conven-
tional collection of ground-water samples for chem-
ical analyses, the quality and cost-effectiveness of
subsurface contamination remediations may be
improved significantly.
Site Characterization Approaches
Tables 14, 15, and 16 provide summaries of the
principal activities, benefits, and shortcomings of
three possible site characterization approaches: con-
ventional; state-of-the-art; and state-of-the-science.
Each activity of the conventional approach can be
accomplished with semi-skilled labor and off-the-
shelf technology. Together with moderate to low
costs, these readily available tools and techniques
are reason enough for perpetuation of the conven-
tional approach - until one notes the shortcomings.
Conventional approaches cannot thoroughly charac-
terize the extent and probable behavior of a sub-
surface contaminant plume; they are, by design, a
compromise between the desire to discover the key
problems at a site and the equal desire to keep
expenses to an absolute minimum.
A comparison of Tables 14, 15, and 16 suggests that
state-of-the-art and state-of-the-science approaches
may be more costly to implement in site character-
izations, but also that the increased value of the
information obtained is likely to save costs because
of dramatic improvements in the technical effective-
ness (e.g., all portions of the zone of contamination
cleansed) and efficiency (e.g., treatment of the mini-
mum volume at the lowest cost) of the site cleanup.
Key management uncertainties regarding the
degree of health threat posed by a site, the selection
of appropriate remedial action technologies, and the
duration and effectiveness of the remediations
should decrease significantly with the implementa-
tion of more sophisticated site characterization
approaches.
The economic benefits of advanced site charac-
terization approaches are illustrated conceptually in
Figure 123. The illustration implies that modest
increases in site characterization expenses (pre-
sumably for more sophisticated data collection and
interpretation efforts) will generate large decreases
in cleanup costs by virtue of greater effectiveness
and efficiency of the remedial design and operation.
In kind, total costs would fall dramatically since
cleanup costs normally comprise the majority of site
expenditures. Maximum return on increased invest-
ments is expected for the state-of-the-art approach
133
-------
and will dimmish as the state-of-the-science ap-
proach is reached because highly specialized equip-
ment and personnel are not widely available.
Testing these conceptual relationships directly is not
possible because an investigation and remediation
cannot be carried out to fruition along each approach
Table 14. Conventional Approach to Site Characterization
Actions Typically Taken
Install several dozen shallow monitoring wells
Sample and analyze numerous times for 129 + priority
pollutants
Define geology primarily by driller's log and cuttings
Evaluate hydrology with water level maps only
Possibly obtain soil and core samples for chemical analyses
Benefits
Rapid screening of problem
Moderate costs involved
Field and lab techniques standardized
Data analysis relatively straightforward
Tentative identification of remedial options possible
Shortcomings
True extent of problem often misunderstood
Selected remedial alternative may not be appropriate
Optimization of remedial actions not possible
Cleanup costs unpredictable and excessive
Verification of compliance uncertain and difficult
Table 15. State-of-the-Art Approach to Site Charac-
terization
Recommended Actions
Install depth-specific well clusters
Sample and analyze for 129 +• priority pollutants initially
Analyze selected contaminants in subsequent samplings
Define geology by extensive coring/split-spoon samplings
Evaluate hydrology with well clusters and geohydraulic tests
Perform limited test on solids (grain size, clay content)
Conduct limited geophysical surveys (resistivity soundings)
Benefits
• Conceptual understanding of problem more complete
• Better prospect for optimization of remedial actions
• Predictability of remediation effectiveness increased
•. Cleanup costs lowered, estimates improved
• Verification of compliance more soundly based
Shortcomings
Characterization costs somewhat higher
Detailed understanding of problem still difficult
Full optimization of remedial actions not likely
Field tests may create secondary problems
Demand for specialists increased
Table 16. State-of-the-Science Approach to Site Charac-
terization
Idealized Approach
• Assume "state-of-the-art-approach" as starting point
• Conduct tracer tests and borehole geophysical surveys
• Determine percent organic carbons, exchange capacity, etc., of
soils and subsurface sediments
• Measure redox potential, pH, dissolved oxygen, etc., of soils
and subsurface sediments
• Evaluate sorption-desorption behavior using select cores
• Assess potential for biotransformation using select cores
Benefits
Thorough conceptual understanding of problem obtained
Full optimization of remedial actions possible
Predictability of remediation effectiveness maximized
Cleanup costs lowered significantly, estimates reliable
Verification of compliance assured
Shortcomings
Characterization cost significantly higher
Few previous field applications of advanced theories
Field and laboratory techniques not yet standardized
Availability of specialized equipment low
Demand for specialists dramatically increased
simultaneously. The best that can be done is to note
the magnitude of changes in perceptions, decisions,
work plans, etc., when advanced techniques are
applied at a ground-water contamination site that
has undergone a conventional level of site charac-
terization. The latter situation is fairly common
because many first attempts at a remedial investi-
gation turn up additional problems or complexities
not suspected when the investigation was budgeted
and implemented, and do not generate consistent or
meaningful information.
Recognizing the need for more technically
sophisticated site characterization efforts is only the
first step toward an improved remedy. Until
recently, implementation of technically sophisti-
cated site characterization approaches was consid-
ered difficult due to the scarcity of skilled labor and
professionals knowledgeable in specialized tech-
niques. Now, however, there is rapid growth in the
number of skilled professionals, as witnessed by the
hundreds of training courses offered annually and
the technical assistance and information transfer
programs within EPA and other Federal and State
agencies. Legally, the passage of the Superfund
Amendments and Reauthorization Act (SARA) gave
EPA and interested parties the opportunity to test
promising technologies at Superfund sites (e.g., the
Superfund Innovative Technology Evaluation (SITE)
program at EPA's Office of Research and
Development).
134
-------
HIGH
CO
O
o
111
LOW
CONVENTIONAL STATE-OF- STATE-OF-
APPROACH THE-ART THE-SCIENCE
Figure 123. Conceptualization of trade-offs of costs in
investigations and cleanups, as a function
of approach used for site characterization.
Illustrative Scenarios
It is helpful to examine possible scenarios that might
result from these different site investigation
approaches. Figure 124 depicts a hypothetical
ground-water and soil contamination site located in
a mixed residential and light industry section of a
town in the Northeast. As illustrated, there are three
major plumes: an acids plume (e.g., from electrolytic
plating operations); a phenols plume (e.g., from a
creosoting operation that used large amounts of
pentachlorophenol); and a volatile organics plume
(e.g., from solvent storage leaks). In addition, on-site
soils are heavily contaminated in one area with
spilled pesticides and in another area with spilled
transformer oils that contain high concentrations of
PCBs.
The hydrogeologic setting for the hypothetical site is
a productive alluvial aquifer that is composed of an
assortment of sands and gravels interfingered with
clay and silt remnants of old streambeds and
floodplains deposits that were continually dissected
and crosscut by a central river as the valley matured
over geologic time. The deeper portion of the
sediments is highly permeable and is the zone most
heavily used for municipal and industrial supply
wells, whereas the shallow portion of the sediments
is only moderately permeable since it contains many
more occurrences of clay and silt lenses. The pre-
dominant ground-water flow direction in the deeper
zone parallels the river (which also is parallel to the
axis of the valley), except in localized areas around
municipal and industrial wellfields. The predom-
inant direction of flow in the shallow zone is
seasonally dependent, with a strong component of
flow toward the river during periods of low flow in
tributary
river
s-^-v^ v
"*» ,'' \
••--.' acids plume
Figure 124. Hypothetical ground-water contamination problem.
135
-------
the river, and a strong component of flow parallel to
the river during periods of high river flow. Strong
downward components of flow carry water from the
shallow zone to the deeper zone throughout
municipal and industrial wellfields, as well as along
the river during periods of high flow. Slight
downward components of flow exist elsewhere due to
local recharge by infiltrating rainwater.
A conventional level of site characterization would
ostensibly define the horizontal extent of the most
mobile/widespread plume, but would provide only a
superficial understanding of variations in the compo-
sition of the sediments. An average coefficient of
permeability (hydraulic conductivity) would be
obtained from review of previously published
geologic reports and assumed to be representative of
the entire aquifer for the purpose of estimating
flowrates.
The kind of cleanup likely to result from a conven-
tional site investigation is illustrated in Figure 125.
The volatile organics plume would be considered the
most important to remediate since it is the most
mobile, and an extraction system would be installed.
Extracted fluids would be air-stripped of volatiles
and then passed through a treatment plant for
removal of non-volatile residues, probably by rela-
tively expensive filtration through granular acti-
vated carbon.
Extraction wells would be placed along the
downgradient boundary of the volatile organic
compound (VOC) plume to withdraw contaminated
ground water. A couple of injection wells would be
placed upgradient and used to return a portion of the
extracted and treated waters to the aquifer. The
remainder of the pumped and treated waters would
be discharged to the tributary under a National
Pollution Discharge Elimination System (NPDES)
permit. Information obtained from the drilling logs
and samples of the monitoring wells could do no more
than position all of the screened sections of .the
remediation wells at the same depth (shallow). The
remediation wellfield would be scheduled to operate
for the amount of time needed to remove a volume of
water somewhat greater than that estimated to
reside within the bounds of the zone of contam-
ination. The latter would amount to perhaps three to
five times the nominal value of contaminated water
and would be based on average retardation values
(found in the scientific literature) for contaminants
found at the site. The PCB-laden soils would be
excavated and sent off to an incinerator or approved
waste treatment and disposal facility. The
decision-makers would have based their approval of
tributary
river
S-A--XV
~-^ •evy-' A
$®®&r- vr.-' H W *&
W&^ , *km* (. •&.. ?. .;;>VK:; "p
:jfe •; > •• . ,'S «;fa-r*., *!*• ,; -;^ ^,a^
!1|::!: ^^|s».rftVv*^ ••>. -|^'::!^
'*•,
Figure 125. Typical conventional cleanup applied to ground-water contamination problem in Figure 124.
136
-------
such a remedy on the presumption that the plume(s)
were adequately defined, or that the true magnitude
of the problem does not differ substantially from
these definitions, save for the possibility of a longer
period of pumpage.
Incorporation of some of the more common state-of-
the-art site investigation techniques, such as pump
tests, installation of vertically separated clusters of
monitoring wells (shallow, intermediate, and deep)
and river stage monitors, and chemical analysis of
sediment and soil samples would likely result in the
kind of remediation illustrated in Figure 126. Since a
detailed understanding of the geology and hydrology
would be obtained, optimal selection of well loca-
tions, wellscreen positions, and flowrates (gallons
per minute) for the remediation wells could be
determined. A special program to recover the acid
plume and neutralize it could be instituted as well as
a special program for the pesticide plume. This
approach would probably lower treatment costs over-
all, despite the need for separate treatment trains for
the different plumes, because substantially lower
amounts of ground water would be treated by expen-
sive carbon filtration.
The extraction wellscreens' positions would become
increasingly deeper as one gets closer to the river
because monitoring well clusters would have indi-
cated that the plume is migrating beneath shallow
accumulations of clays and silts to the deeper, more
permeable sediments. Approximately two-thirds of
the extracted and treated ground water would be
reinjected through injection wellscreens positioned
deep to avoid diminishing the effectiveness of nearby
extraction wells. As in the conventionally based
remedy, the remediation wellfield would be sched-
uled to operate for the amount of time needed to
remove a volume of water based on average contam-
inant retardation values and the volume of ground
water residing in the zone of contamination. The
detailed geologic and hydrologic information ac-
quired, however, would result in an expectation of a
more rapid cleansing of specific portions of the zone
of contamination. The decision-makers would have
based their approval of this remedy on the presump-
tion that the remediation is optimized to the point of
providing the most effective cleanup, though the
efficiency of the remediation may be less than
optimal.
If all state-of-the-art investigation tools were used at
the site, there would be an opportunity to evaluate
the desirability of using a subsurface barrier wall to
enhance remediation efforts (Figure 127). The wall
would not be expected to entomb the plumes, but
tributary
river
pesticide area
(pump & treat w/ GAG)
treatment
facility
EW6* %l phenols plume
•
VOC plume
(pump & air strip)
acids plume
(pump & neutralize)
•• . ;•;:;:;:;; •'••:.-..' .:/screen settings of.-
' ':-::':'-'.- .' ':•:] reclamation .wells :
Figure 126. Moderate state-of-the-art cleanup applied to ground-water contamination problem in Figure 124.
137
-------
would limit pumping to contaminated fluids (rather
than having the extracted waters diluted with fresh
waters available to the extraction wells, as was true
of the two previous approaches). The volume pumped
would be lower because the barrier wall would
increase the drawdown at each well by hydraulic
interference effects, thereby maintaining the same
effective hydrodynamic control with lesser pumpage
(note the lower values in the sets of parentheses at
each well in Figure 127, given in gallons per
minute). Treatment costs would decrease too,
because the waters pumped would contain higher
concentrations of contaminants (treatment effi-
ciencies normally fall with decreasing concentra-
tions). Soil washing techniques would be used on the
pesticide-contaminated area to minimize future
source releases to ground water.
The efficiency and effectiveness of the remediation
would appear to be optimal, but that is a perception
based on the presumption that contaminants are
readily released. Given the potential limitations to
pump-and-treat remediations discussed in earlier
sections of this document, however, it is doubtful
that this advanced state-of-the-art site investigation
precludes further improvement. Chemical and
biological peculiarities must be given as much
attention as the site geology and hydrology. The use
of average retardation values from the literature
infers that additional improvements in effectiveness
and efficiency can be garnered by detailed evaluation
of contaminant retardation at this site. Likewise,
detailed examination of the potential for biotrans-
formation would be expected to provide additional
effectiveness and efficiency.
At the state-of-the-science level of site character-
ization, tracer tests could be undertaken that would
provide good information on the potential for
diffusive restrictions in low permeability sediments
and on anisotropic biases in the flow regime.
Sorption behavior of the VOCs could be evaluated in
part by determination of the total organic carbon
contents of the subsurface sediments. Similarly, the
cation exchange capacities of subsurface sediment
samples could be determined to obtain estimates of
release rates and mobilities of toxic metals. The
stabilities of various possible forms of elements and
compounds could be evaluated with measurements of
pH, redox potential, and dissolved oxygen. Finally, if
state-of-the-science findings regarding potential bio-
transformations were used, it might be possible to
effect in situ degradation of the phenols plume and
remove volatile residues (Figure 128).
tributary
V
river
'/0EW7
', (100' pesticide area —
\ (wash soils, pump
\ & treat w/ GAG)
"Sx; » * Y Yt Y,
"?' ,' PCB soils
^ " ~ . (remove)
-. ."-N ^
X X.
EW6
(150)
phenols plume
^V /
-: _Ews •m •'
X»(150i..- —--—f-7.!). /
'-—-*" >^ \X
treatment
facility
IW2 ^-^/gf.
(iso) »s fh
(150)/^''J
S
shallow subsurface wall
(to maximize pumpage of plume)
acids plume
(pump & neutralize)
:BWt
4-'.
KEW6-'
VST
? -wwf^«> -. -^;:
Ffgura 127. Advanced state-of-the-art cleanup applied to ground-water contamination problem in Figure 124.
138
-------
tributary
river
pesticide area
(wash soils, pump
& treat w/GAG)
treatment •
facility
.' PCBsoilsl
(remove)
phenols plume
(bioreclaim)
shallow subsurface wall
(to maximize pumpage
of plume)
(pump & air strip
majority, bioreclaim
residuals)
acids plume
(pump & neutralize)
Figure 128. State-of-the-science cleanup applied to ground-water contamination problem in Figure 124.
Performance Evaluations of
Remediations
Large expenditures are made each year to prepare
for and operate remediations of ground-water con-
tamination. Regulatory responsibilities require that
adequate oversight of these remediations be made
possible by structuring appropriate compliance
criteria for monitoring wells. The oversight efforts
are nominally directed at answering the question,
"What can be done to show whether or not a
remediation is generating the desired contamination
control?" Recently, other questions have developed
because of the realization that pump-and-treat
remediations do not function as well as has been
presumed. Such questions include: "What can be
done to determine whether the remediation will
meet its timelines?" and "What can be done to
determine whether the remediation will stay in
budget?"
Conventional wisdom states that these questions can
be answered by the use of sophisticated data analysis
tools, such as computerized mathematical models of
ground-water flow and contaminant transport. Com-
puter models can indeed be used to make predictions
about future performance, but such predictions are
highly dependent on the quality and completeness of
the field and laboratory data. The latter is just as
true for models evaluating pump-and-treat remedi-
ations, in contrast to the common belief that an
accurate performance evaluation can be made simply
by comparing data obtained from monitoring wells
during remediation to the data generated prior to the
.onset of remediation. Historical trends of contam-
inant levels at local monitoring wells are rendered
useless by the extraction and injection wells used in
pump-and-treat remediations. This is a consequence
of the fact that the extraction and injection wells
produce complex flow patterns locally, where pre-
viously there were comparatively simple flow
patterns.
Complex ground-water flow patterns present great
technical challenges in terms of characterization and
manipulation (management) of the associated
contaminant transport pathways. In Figure 129, for
example, waters moving along the flowline that
proceeds directly into a pumping well from up-
gradient are moving the most rapidly, whereas those
waters lying at the lateral limits of the capture zone
(indicated by the bold curved line in Figure 129)
move much more slowly. One result is that certain
parts of the aquifer are flushed quite well and others
are remediated relatively poorly. Another result is
that those previously uncontaminated portion^ of the
139
-------
MODERATE
FAST
Figure 129.
MODERATE
Flowline pattern generated by an extraction
well.
aquifer that form the peripheral bounds of the
contaminant plume may become contaminated by
the operation of an extraction well that is located too
close to the plume boundary. This occurs because the
flowline pattern extends downgradient of the well.
The latter is not a trivial situation avoided without
repercussions by simply locating the extraction well
far enough inside the plume boundary that its
flowline pattern does not extend beyond the down-
gradient edge of the plume, because doing so results
in very poor cleansing of the aquifer between the
location of the extraction well and the downgradient
plume boundary.
It is not possible to determine precisely where the
various flowlines generated by a pump-and-treat
remediation are located unless detailed field evalu-
ations are made during remediation. Neither con-
taminant nor velocity distributions are constant
throughout the zone of action (that portion of an
aquifer actively manipulated by the pumping wells).
Consequently, more data must be generated during
the remediation (especially inside the boundaries of
the contamination plume) than were generated
during the entire remedial investigation/feasibility
study process at a site, and interpretations must be
made of those data that require much more sophisti-
cated tools. Indeed, it might be successfully argued
that in most settings, monitoring well data collected
during remediation are useless unless a mathe-
matical model is used to organize and analyze the
data.
Decisions regarding the frequency and density of
chemical samplings must consider the detailed flow-
paths generated by the remediation wellfield, and
include changes in contaminant concentrations
resulting from variations in the influences of trans-
port processes along those flowpaths. The need to
reposition extraction wells occasionally to remediate
portions of the contaminated zone previously subject
to slow flowlines means that the chemical samplings
may generate results that are not easily understood.
It also means that the chemical compliance points
may have to be moved during the course of a
remediation.
Nor are evaluations of the hydrodynamic per-
formance of remediation wellfields easily accom-
plished. For example, an inward hydraulic gradient
is usually required to be maintained at the periphery
of a contaminant plume undergoing remediation by
use of a pump-and-treat wellfield. This requirement
is imposed to ensure that no portion of the plume is
free to migrate away from the zone of action. To
assess this performance adequately, the hydraulic
gradient must be measured accurately in three
dimensions between each pair of adjacent pumping
or injection wells. The design of an array of
piezometers (small diameter wells with very short
screened intervals, used to measure the pressure
head of selected positions in an aquifer) for this
purpose can be difficult. Two points define a line and
three points define a planar surface, but many more
are needed to define the convoluted water-table
surface that develops between adjacent pumping or
injection wells. Not only are there velocity divides in
the horizontal dimension near active wells, but in
the vertical dimension, too, because the pressure
influence of each well extends to only a limited depth
in practical terms.
Innovations in Pump-and-Treat
Remediations
One of the promising innovations in pump-and-treat
remediations is intermittent operation or pulsed
pumping of a remediation wellfield. Pulsed operation
of hydraulic systems is the cycling of extraction or
injection wells on and off in "active" and "resting
phases" (Figure 130). The resting phase of a pulsed-,
pumping operation can allow sufficient time for
contaminants to diffuse out of low permeability, zones
and into adjacent high permeability zones, until
maximum concentrations are achieved in the higher
1 'a '3 '4 's 'e '? 'a
Figure 130. Reduction of residual contaminant mass by
pulsed pumping.
140
-------
permeability zones. For sdrbed contaminants and
NAPL residuals, sufficient time can be allowed for
equilibrium concentrations to be reached in local
ground water. Subsequent to each resting phase, the
active phase of the pulsed-pumping cycle removes
the minimum volume of contaminated ground water,
at the maximum possible concentrations, for the
most efficient treatment. By occasionally cycling
only select wells, these wells' stagnation zones may
be brought into active flowpaths and remediated.
Pulsed operation of remediation wellfields incurs
certain additional costs and concerns that must be
compared with its advantages for site-specific appli-
cations^" During the rest phase of pulsed-pumping
cycles, peripheral gradient control may be needed to
ensure adequate hydrodynamic control of the plume;
in an ideal situation, peripheral gradient control
would be unnecessary. This might be the case where
there are no active wells, major streams, or other
significant hydraulic stresses nearby to influence the
contaminant plume while the remedial action
wellfield is in the resting phase. The plume would
migrate only a few feet during the tens to hundreds
of hours that the system was at rest, and that
movement would be rapidly recovered by the much
higher'flow velocities back toward the extraction
wells during the active phase.
When significant hydraulic stresses are nearby,
however, plume movement during the resting phase
may be unacceptable. Irrigation or water-supply
pumpage, for example, might cause plume
movement on the order of several tens of feet per day.
It then might be impossible to recover the lost
portion of the plume when the active phase of the
pulsed-pumping cycle commences. In such cases,
peripheral gradient control during the resting phase
would be essential. If adequate storage capacity is
available, it may be possible to provide gradient
control in the resting phase by injection of treated
waters downgradient of the remediation wellfield.
Regardless of the mechanics of the compensating
actions, their capital and operating expenses must be
added to those of the primary remediation wellfield.
Pump-and-treat remediations currently are under-
way that incorporate some of the principles of pulsed
pumping. For instance, pumpage from contaminated
bedrock aquifers and other low permeability forma-
tions results in intermittent wellfield operations by
default; the wells are pumped dry even at low flow
rates. In such cases, the wells are operated on
demand with the help of fluid-level sensors that
trigger the onset and cessation of pumpage. This
simultaneously accomplishes the goal of pumping
ground water only after it has reached chemical
equilibrium, since equilibrium occurs on the same
time frame as the fluid recharge event (both are
diffusively restricted). In settings of moderate to
high permeability, the onset and cessation of
pumpage could be keyed to contaminant
concentration levels in the pumped water, independ-
ent of flow changes required to maintain proper
hydrodynamic/gradient control. As indicated in the
discussion of pulsed pumping, this may be acceptable
(pose no unreasonable risk) in circumstances where
the contaminant plume would not be subject to
substantial movement in the absence of pumpage.
Other strategies for improvement of the performance
of pump-and-treat remediations include:
1. Flow scheduling of wellfield operations to satisfy
simultaneously hydrodynamic/gradient control
and contaminant concentration trends or other
performance criteria.
2. Physical repositioning of extraction wells to
effect major flowline/transport pathway altera-
tions.
3. Integration of wellfield operations with other
subsurface technologies (e.g., barrier walls to
limit plume transport and minimize pumping of
fresh water, or infiltration ponds to maintain
saturated flow conditions for flushing contam-
inants from (normally) unsaturated soils and
sediments).
The first of these alone would allow for flushing of
stagnant zones by occasionally turning off individual
pumps, but the flushing could not be done as
efficiently as repositioning or adding pumping wells
(the second means of improvement). The first and
second approaches differ in effects, however, because
repositioning or adding wells requires access for
drilling and necessarily precludes capping of the site
until after completion of the pump-and-treat opera-
tions. The third improvement approach, combining
pump-and-treat with subsurface barrier walls,
trenching, or in situ techniques (all of which may
occur at any time during remediation), also may
require postponement of capping until after comple-
tion of the remediation.
The latter strategy raises latent fears of lack of
control of the contaminant source, which is almost
always mitigated by isolation of the contaminated
soils and subsoils that remain long after man-made
containers are removed from the typical site.
Fortunately, vacuum extraction of contaminanted
air/vapor from soils and subsoils has recently
emerged as a potentially effective means of removing
VOCs, steam flooding is being evaluated for removal
of the more retarded organics, and in situ chemical
fixation techniques are being tested for the isolation
of metals wastes. Vacuum extraction is capable of
removing several pounds of VOGs per day (since the
VOCs readily volatilize into the soil gas/vapor),
141
-------
whereas air stripping of VOCs from comparable
volumes of contaminated ground water typically
results in the removal of only a few grams of VOCs
per day (because VOCs are so poorly soluble in
water). Similarly, steam flooding can be an econom-
ically attractive means of concentrating contami-
nant residuals as a front leading the injected body of
steam. Regardless of the efficacy of vacuum extrac-
tion, steam flooding, or chemical fixation in terms of
permanent and complete remediation of the contami-
nation in the unsaturated zone each have excellent
potential for control of fluid and contaminant
movement in the unsaturated zone and should be
considered as potentially significant additions to the
list of source control options. In addition, soils
engineering and landscape maintenance techniques
can minimize infiltration of rainwater in the absence
of a multilayer RCRA-style cap.
In terms of performance evaluation of a remediation,
the presence of a multilayer RCRA-styled cap poses
major limitations. The periodic removal of core
samples of subsurface solids from the body of the
plume and the source zone, with subsequent extrac-
tion of the chemical residues on the solids, is the only
direct means of evaluating the true magnitude of the
residuals and their depletion rate. Since this must be
done periodically, capping should be postponed until
closure of the site.
If capping can be postponed or forgone, there will be
great flexibility for management of pump-and-treat
remediations which can improve effectiveness and
lower costs. Also, the soils and subsoils can be
cleansed of contamination without waiting in isola-
tion for eventual breakdown of a cap. Given the
parallel theme of SARA - true remediation, not just
stabilized problems - innovative pump-and-treat re-
mediations and source removal techniques may be
the most economical and responsible choices for
remediations.
Managerial Considerations in Using
Transport and Fate Models
Effective Communications
One of the principal problems underlying the
continuing difficulties in transferring technical
information about transport and fate issues is the
poor level of communications between specialists and
decision-makers that often results in poorly focused
remediation activities. Part of this problem occurs
because specialists and decision-makers have dif-
ferent perceptions of their roles and the situations
they face. The problem of effective communication is
not easy to solve. Questions presented in Tables 17,
18, and 19 can be used to stimulate more effective
dialogues and may provide insights to presenting
material for public consumption, both in terms of
clarity and honest appraisals of the costs and
limitations that must be accepted by the public.
Table 17. Screening Level Questions to Help Focus
Ground-water Contamination Assessments
General Problem Definition
• What are'the key issues: quality, quantity, or both?
• What are the controlling geologic, hydralogic, chemical, and
biological features?
• Are there reliable data (proper field scale, quality controlled,
etc.) for preliminary assessments? ,„_„
• Do the model(s) needed for appropriate simulations exist?
Initial Responses Needed
• What is the time-frame for action (imminent or long-term)?
• What actions, if taken now, can significantly delay or
minimize the projected impacts?
• To what degree can mathematical simulations yield
meaningful results for the action alternatives, given available
data?
• What other techniques or information (generic models, past
experience, etc.) would be useful for initial estimates?
Strategies for Further Study
• Are the critical data gaps identified; if not, how well can
specific data needs be determined?
• What are the trade-offs between additional data and
increased certainty of the assessments?
• How much additional manpower and resources are necessary
to improve mathematical modeling efforts?
• How long will it take to produce useful simulations, including
quality control and error-estimation efforts?
Technical Support
The return on investments made in using transport
and fate models rests principally with the training
and experience of the technical support staff
applying the model to a problem and on the degree of
communication between those persons and manage-
ment. In discussing the potential uses of computer
modeling for ground-water protection efforts, Faust,
et al. (1981) noted that the final worth of modeling
applications depends on the people who apply the
models. Managers should be aware that specialized
training and experience is necessary to develop and
apply mathematical models, and relatively few
technical support staff can be expected to have such
skills (van der Heijde, et al., 1985). This is due in
part to the need for the modeler to have familiarity
with a number of scientific disciplines so that the
model is structured faithfully to simulate real-world
situations.
142
-------
Table 18. Conceptualization Question* to Help Focus
Ground-water Contamination Assessments
Field Techniques and Data Production
• Are the installation and sampling techniques to be used
accepted? innovative? controversial?
• Where are the weak spots in the assessment, and can these
be further minimized or eliminated?
• What are the limitations of field tests that estimate the natural
processes parameters of this problem?
Model Input Parameters and Boundary Conditions
• How reliable are the estimates of the input parameters; are
they quantified within accepted statistical bounds?
• What are the boundary conditions, and why are they
appropriate to this problem?
• Have the initial conditions with which the model is calibrated
been checked for accuracy and consistency?
• Are the spatial grid design(s) and time-steps of the model
optimized for this problem?
Model Quality Control and Error Estimation
• Have these models been mathematically validated against
other solutions to this kind of problem?
• Has anyone field-verified these models before, by direct
applications or simulation of controlled experiments?
• How do these models compare with others in terms of
computational efficiency, ease of use, or modification?
• What special measures are being taken to estimate the
overall errors of the simulation?
What levels of training and experience are necessary
to apply mathematical models properly? Are "Ren-
naissance specialists" needed or can interdisci-
plinary teams be effective? The answers to these
questions are not clear-cut, but the more informed an
individual is, the more effective he or she can be. It is
doubtful that any individual can master each
discipline with the same depth of understanding as
specialists in those fields, but a working knowledge
of many disciplines is necessary so that appropriate
questions may be put to specialists, and some sense
of integration of the various disciplines can evolve.
In practice, this means that modelers should be
involved in continuing education efforts with the
support of management. The benefits to be gained
are tremendous, and the costs of not doing so may be
equally large.
Managers and technical support staff alike should
appreciate the difficulty in explaining the results of
complicated models to non-technical audiences such
as in public meetings and courts of law. Many
scientists find it difficult to discuss the details of
their labors without the convenience of their scien-
tific jargon. Some of the more useful means of
overcoming this limitation involve the production of
highly simplified audio-visual aids.
Table 19. Sociopolitical Questions to Help Focus Ground-
water Contamination Assessments
Demographic Considerations
• Is there a larger population endangered by the problem than
we are able to provide sufficient responses to?
• Is it possible to present this assessment in both non-technical
and technical formats to reach all audiences?
• What role can modeling play in public information efforts
(e.g., effective graphics)?
• ;How prepared are we to respond to criticism of this
assessment (e.g., supportive materials)?
Political Constraints
• Are there non-technical barriers to the techniques to be used
to produce this assessment, such as "tainted by association"
with a controversy elsewhere?
• *Do we have the cooperation of all involved parties in
obtaining the necessary data and implementing solutions?
• Are similar technical efforts for this problem being undertaken
by friend or foe?
• Can the results of the assessment be turned against us; are
the results ambiguous or equivocal?
Legal Concerns
• Will these activities meet all regulations?
• If we are dependent on others for key inputs, how do we
recoup losses stemming from possible non-performance?
• What liabilities are incurred for projections arising from poor
data, misinterpretations, or models used?
• Do any of the issues to be addressed by this assessment
require the advice of attorneys?
Potential Liabilities
Some of the liabilities in using transport and fate
models relate to the degree to which predictive
models are used in permitting or banning specific
practices or products. If a model is incapable of
treating specific applications properly, substantially
incorrect decisions may result. Depending on the
application, unacceptable environmental effects may
begin to accumulate long before the nature of the
problem is recognized. Conversely, unjustified
restrictions may be imposed on the regulated
community. Inappropriate or inadequate models also
may cause the re-opening clause of a negotiated
settlement agreement to be invoked when, for
instance, compliance requirements that were guided
by the predicted plume behavior generated by the
model are not met.
Certain liabilities relate to the use of proprietary
codes in legal settings, where the inner workings of a
model may be subject to disclosure in the interests of
justice. The desire for confidentiality by the model
developer would likely be subordinate to the public
right to full disclosure of actions predicated on
modeling results. The mechanisms for protection of
143
-------
proprietary rights typically do not extend beyond
extracted promises of confidentiality by reviewers or
other interested parties. Hence, a developer of
proprietary codes still risks exposure of innovative
techniques.
Other liabilities may arise as the result of mis-
applications of transport and fate models, or
applications of models later found to be faulty.
Frequently, the choices of boundary and initial
conditions for a given application are hotly con-
tested; misapplications of this kind are undoubtedly
responsible for many of the reservations expressed
by would-be model users. Also, many times in the
past, a well-used and highly regarded model code
was found to contain errors or lack the ability to
faithfully simulate certain situations for which it
was widely presumed to be applicable (El-Kadi,
1988). The best way to minimize these liabilities is to
adopt strict quality control procedures for each
application.
Economic Considerations
The nominal costs of the support staff, computing
facilities, and specialized graphics production equip-
ment associated with transport and fate modeling
efforts can be high. In addition, quality control
activities can result in substantial costs. The deter-
mining factor in controlling these costs is the degree
to which a manager must be certain of the charac-
teristics of the model and the validity of its output.
As a general rule, costs are greatest for personnel,
moderate for hardware, and minimal for software.
The exception to this ordering relates to the com-
bination of software and hardware purchased. An
optimally outfitted business computer (e.g., VAX
11/785 or IBM 3031) costs higher than $100,000, but
can rapidly pay for itself in terms of dramatically
increased speed and computational power. A well-
complemented personal computer (e.g., Compaq
Deskpro 386 or Macintosh II) may cost $10,000 with
moderate accessories, but the significantly slower
speed and limited computational power may infer
hidden costs in terms of the inability to perform
specific tasks. For example, highly desirable statis-
tical packages like SAS and SPSS are only available
with reduced capabilities or altogether unavailable
for personal computers. Many of the most sophisti-
cated mathematical models are only available in
their fully capable form on business computers.
Figure 131 gives a brief comparison of typical costs
for software at different levels of computing power.
Obviously, the software for less capable computers is
cheaper, but the programs are not equivalent;
therefore, managers need to consider the appropri-
ateness of a chosen level of computer power. If the
decisions to be made will be based on very little data,
it may not make sense to insist on the most elegant
software and hardware. If the intended use involves
substantial amounts of data and sophisticated
analyses are desired, it would be unwise to opt for the
least expensive combination.
300
Mainframe Fortran PC Fortran PC BASIC
Ground-Water Modeling Software Categories
minimum cost
maximum cost
Figure 131.
Average price per category for ground-
water models from the international
ground-water modeling center.
There seems to be an increasing drive away from
both ends of the spectrum of computing power and
toward its middle; that is, the use of. powerful
personal computers is increasing rapidly, whereas
the use of small programmable calculators and large
business computers is declining. In part, this stems
from the significant improvements in the computing
power and quality of printed outputs obtainable in
recent years from personal computers.; Also, the
telecommunications capabilities of personal com-
puters now commonly include emulation of the
interactive terminals of large business computers so;
that vast computational power can be accessed and
the results retrieved with no more than a phone call.
Recently, many of the mathematical models and data
packages have been 'down-sized' from mainframe
computers to personal computers and more are now
being written directly for this market.
144
-------
Figure 132 provides some idea of the costs of
available software and hardware for personal com-
puters. Table 20 lists recent salary ranges and
desired background for the technical support staff
needed to operate such systems. Figure 133 attempts
to place all of the nominal costs of subsurface
contaminant transport modeling in perspective.
Table 20. Desired Backgrounds and Salary Ranges Adver-
tised for Positions' Requiring Ground-water
Modeling
2.00O
g 1.600
CO
g 1,000
oZ
CO 500
1
ft
-
Jin
•-
Jl
•nil
-
1
•
Jinn
2 3 4 5 6 7 8 9 10 11 12 13 14 15
Vendors of Ground-Water Models
| I
1. International Ground
Water Modeling Ctr.
2. Computapipe Co.
3. Data Services, Inc.
4. GeoTrans, Inc.
5. Hydrosoft, Inc.
6. In Situ, Inc.
7. IrriscoCo.
8. Koch and Assoc.
9. KRS Enterprises, Inc.
10. Michael P. SpinksCo.
11. RockWare, Inc.
12. Solutech Corp.
13. T.A. Prickett & Assoc.
14. James S. Ulrick Co.
15. Watershed Research, Inc.
Figure 132. Price ranges for IBM-PC ground-water
models available from various sources.
The technical considerations discussed in previous
sections indicate that the desired accuracy of the
modeling effort directly affects the total costs of
mathematical simulations. Thus, managers will
want to determine the incremental benefits gained
by increased expenditures, especially for more in-
volved mathematical modeling efforts. While many
economic theories exist for determining these bene-
fits, the most straightforward of these are the cost-
benefit approaches commonly used to evaluate the
economic desirability of water resource projects.
There are two generalized approaches commonly
used: the Benefit/Cost Ratio method and the Net
Benefit method.
The Benefit/Cost (B/C) Ratio method involves
tallying the economic value of all benefits and
dividing that sum by the total cost involved in gener-
ating those benefits. A ratio greater than one is
required for the project to be considered viable,
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Position Title:
Salary Offered:
Desired Background:
Hydrogeologist (Argonne Natl. Lab.)
$31,619-$48,876
Familiarity and experience in field testing
and monitoring of ground-water flow and
the use of numerical models
Hydrologic Modeler (University of Wis-
consin)
$23,000-$25,000
Primary strength in application of
numerical models to ground-water flow
and chemical transport; strong chemical
background
Soil Scientist (USEPA)
$31,619-$41,105
Knowledge of: (1) soil physics; (2)
processes governing transport and fate
of chemical and biological species (3)
math, statistics, and geostatistics; and
the ability to develop computer codes
Ground-water Hydrologist (Inyo County
Water Dept, CA)
Starting up to $32,000
At least three years experience including ,
field work, surface/ground-water resource
evaluations, environmental assessments,
flow modeling; FORTRAN
Hydrogeologist (S.W. Texas State
University)
$24,444-$30,096
Academic training in hydrogeology, min.
2 years experience, knowledge of
limestone aquifers and computer
operations
Geochemist {U.S. Nuclear Regulatory
Com.)
$21,170-$41,105
Knowledge of solute and radionuclide
transport, including speciation,
attenuation (sorption), numerical
modeling.
Hydrogeol./Civil Engr. (typical consulting
firm)
"commensurate with experience"
Strong background in applied ground-
water flow and contaminant transport
modeling, knowledge of Federal/State
regulations
though there may be sociopolitical reasons for
proceeding with projects that do not meet this
criterion. An example would be a new project that
has gained considerable social or political momen-
tum, but which begins to exceed initial cost
estimates. Not proceeding or substantially altering
the work may be economically wise; however, such a
decision may be viewed as a breach of faith by the
public. Regardless of how this kind of situation
evolves, it is not uncommon for certain costs to be
145
-------
Compaq 386
& Accessories
w/Financing
($15K)
Ground-Water
Models ($5K)
Benefits -
25% Salary
Service,
Expendables,
and Optional
Software ($10K)
Salary: 1.0 FTE
- $40K/yr
Total Cost: $280K
Ground-Water
Models ($15K)
Benefits -
25% Salary
VAX 11/785
& Accessories
w/Financing
($125K)
Service,
Expendables,
and Optional
Software ($25K)
Salary: 1.0 FTE
- $40K/yr
Total Cost: $415K
Figure 133. Costs of sustaining ground-water modeling capabilities at two different computing levels for a five-year period.
forgiven or subsidized, muddying the picture for
incremental benefits or trade-off analyses.
The Net Benefit method involves determining the
arithmetic difference of the total benefits and total
costs. The obvious criterion in this method is that the
proposed work result in a situation where total
benefits exceed total costs. This approach is most
often adopted by profit-making enterprises because
they seek to maximize the difference, which is their
source of income. The B/C Ratio method, by contrast,
has long been used by government agencies and
other non-profit organizations because they seek to
show the simple viability of their efforts, irrespective
of the costs involved.
In a very real sense, these two general economic
assessment methods stem from different philoso-
phies, yet they share many common difficulties and
limitations. For example, there is a need to predict
the present worth of future costs and to amortize
benefits over the life of a project. The mechanics of
such calculations are well known, but they necessar-
ily involve substantial uncertainties. The present
146
-------
worth of a series of equal payments for equipment or
software can be computed by:
P = A * [((1 + i)n - l)/(i * (1 + i)n)]
where P is the present worth, A is the series payment
each interest period, i is the interest rate per period,
and n is the number of interest periods (White, et al.,
1984). Note, however, that the interest rate must be
estimated. A small difference in the interest rate
results in tremendous differences in the present
worth estimate due to the exponential nature of the
equation.
Managers also may compute the future worth of a
present investment, calculate the percentage of
worth annually acquired through single payments or
serial investments, and so on. One should be aware
that these methods of calculating costs belong to the
general family of single-objective, or mutually exclu-
sive alternative analyses which assume that the cost
of two actions is obtained by simple addition of their
singly computed costs. In other words, the efforts
evaluated are presumed to have no interactions; for
some aspects of ground-water modeling efforts, this
assumption may not be valid. For instance, one may
not be able to specify software and hardware costs
independently. In addition, these methods rely on
the expected value concept, wherein the expected
value of an alternative is viewed as the single
product of its effects and the probability of their
occurrence. This means that high-risk, low-proba-
bility alternatives and low-risk, high-probability
alternatives have the same expected value.
To overcome these difficulties, methods can be used
which incorporate functional dependencies between
various alternatives and do not rely on the expected
value concept, e.g., the multi-objective decision
theories (Asbeck and Haimes, 1984; Haimes and
Hall, 1974; Haimes, 1981). A conceivable use would
be the estimation of lowered health risks associated
with various remedial action alternatives at a haz-
ardous waste site. In such a case, the output of a
contaminant transport model would be used to
provide certain inputs (i.e., contaminant concentra-
tions and transport velocities) to a health effects
model, and that model would produce the inputs
(e.g., probability of additional cancers per level of
contaminant) for the multi-objective decision model.
The primary difficulty with multi-objective ap-
proaches is estimating the probabilities of each
alternative so that the objectives which are to be
satisfied may be ranked in order of importance. A
related difficulty is the need to specify the functional
form of the inputs (e.g., the population distribution
function of pumpage rates or contaminant levels).
Historical records about the inputs may be too
insufficient to allow their functional forms to be
determined.
Another problem compounding the cost-benefit
analysis of mathematical modeling efforts relates to
the need to place an economic value on intangibles.
For example, the increased productivity a manager
might expect as a result of rapid machine calcu-
lations replacing hand calculations may not be as
definable in terms of the improved quality of
judgments made as it is in terms of time released for
other duties. Similarly, the estimation of improved
ground-water quality protection benefits may
necessitate some valuation of human life and suffer-
ing. Hence, there is often room for considerable
adjustment of the values of costs and benefits. This
flexibility can be used inappropriately to improve
otherwise unsatisfactory economic evaluations. For
instance, Lehr (1986) offers a scathing indictment of
the Tennessee Valley Authority for conducting
hydroelectric projects which have "incredibly large
costs" and "negative cost benefit ratios."
Finally, some costs and benefits may be incorrectly
evaluated because they are based on probabilistic
data, a fact which goes unrecognized. For instance,
the key parameters affecting ground-water computa-
tions (e.g., hydraulic conductivity) are only known
within an order-of-magnitude due to data collection
limitations. In these situations, great caution must
be exercised.
Summary
This chapter described the present and future status
of ground-water contamination assessments and the
large difference between what may be known in a
theoretical context and what is put into practice.
This difference exists because field methods used to
characterize important natural process parameters
are still relatively crude and there remains the
unwarranted perception that mathematical models
can estimate these important parameters accurately
with small amounts of data.
Historically, decision-makers have been reluctant to
fund state-of-the-art site characterization ap-
proaches. Such approaches are more frequently
being recognized as the appropriate means by which
to design, implement, and complete the most effec-
tive and efficient remediations of ground-water con-
tamination. Mathematical models can be used to
gain insights to potential behavior of a plume and to
test hypotheses about conceptualizations, so as to
generate better understandings of important phys-
ical, chemical, and biological processes which affect
specific ground-water problems — but only if
adequate data are available.
147
-------
Pump-and-fcreat remediations are far more
complicated than previously thought. The variations
in ground-water flow velocities and directions that
are imposed on natural systems by remediation
wellfields tremendously complicate attempts to
evaluate the progress of pump-and-treat remedi-
ations, in part because of the tortuosity of the
flowlines that are generated, and because historical
trends of contaminant concentrations at monitoring
wells are rendered useless for further predictions and
evaluations. Just as it is improbable that a proper
understanding of the true extent of the contam-
ination problem can be obtained unless sophisticated
site characterization approaches are utilized, it is not
possible to optimize the effectiveness and efficiency
of a pump-and-treat remediation unless the geologic,
hydro logic, chemical, and biological complexities of a
site are defined adequately. While it is evident that
no guarantee can be given that a remediation will
indeed be completely effective and optimally efficient
by virtue of the use of state-of-the-art and state-of-
the-science techniques, it is equally evident that
their use ensures that the best remediation practical
can be achieved.
The use of mathematical models in the decision-
making process means that the user will inevitably
incur certain liabilities. Anticipation of problem
areas and some sensitivity to the possible misuses of
models will greatly minimize potential liabilities.
Rigorous quality control programs also will achieve
the same goal. A number of direct and indirect costs
attend the use of models, not the least of which
involve efforts to obtain and retain specialized
experts. There will always be a significant degree of
professional judgment called for in ground-water
contamination assessments, and, judging from the
frequency with which significant errors are intro-
duced by poor field work, this is an area that needs
much attention. More strict licensing of engineers or
scientists engaged in this kind of work will probably
do little to improve the situation. Rather, much more
effective communication between specialists and
decision-makers, and their communication with the
public is needed. Bringing transport and fate issues
out of the research community and into the political
arena, and describing and addressing the problem of
subsurface contamination is - with all its costs and
technical limitations - the real solution.
References
Asbeck, E. and Y. Y. Haimes, 1984. "The
Multiobjective Risk Method." Large Scale Systems,
Vol. 13, No. 38.
El-Kadi, A., 1988. "Applying .the USGS Mass
Transport Model (MOC) to Remedial Actions by
Recovery Wells." Groundwater, Vol. 26, No. 3.
Faust, C. R., L. R. Silka, and J. W. Mercer, 1981.
."Computer Modeling and Ground-water
Protection." Groundwater, Vol. 19, No. 4.
Haimes, Y. Y., 1981. Risk/Benefit Analysis in Water
Resources Planning and Management. Plenum
Publishers, New York.
Haimes, Y. Y. and W. A. Hall, 1974. "Multiobjectives
in Water Resources Systems Analysis: the
Surrogate Worth Trade-off Method." Water
Resources Research, Vol. 10.
Lehr, J. H., 1986. "The Myth of TVA." Groundwater,
Vol. 24, No. 1.
van der Heijdej P. K. M., et al., 1985. Groundwater
Management: The Use of Numerical Models, 2nd
Edition, AGU Water Resources Monograph No. 5,
American Geophysical Union, Washington, DC.
White, J. A., M. H. Agee, and K. E. Case, 1984.
Principles of Engineering Economic Analysis, 2nd
Edition. John Wiley and Sons, New York.
•C.S. GOVERNMENT PRINTING OFFICE: 1994-550-001/00175
148
-------
-------
-------
-------
TJ O
CD ^
Q) O
> mi
8
Q
0>
00
§
NJ
«I* §2?
g Q) O • 0) Q)
2 n c rr o «
_ IT o Q) J CD
2.0° i.°3
OO
O
M
vo
Si
1 S 3
s ^ ^
? 2.
m
------- |