Sie sind auf Seite 1von 75

United States Office of Research and EPA/625/R-95/005

Environmental Protection Development July 1996

Agency Washington DC 20460

Ground-Water Remediation
A Guide for Decision Makers
and Practitioners
1. Introduction to Pump-and-Treat Remediation
Pump-and-treat is one of the most widely used What is involved in smart application of
ground-water remediation technologies. Conven- the pump-and-treat approach?
tional pump-and-treat methods involve pumping What are tailing and rebound, and how can
contaminated water to the surface for treatment. they be anticipated?
This guide, however, uses the term pump and treat
in a broad sense to include any system where What are the recommended methods for
withdrawal from or injection into ground water is meeting the challenges of effective hydraulic
part of a remediation strategy. Variations and containment?
enhancements of conventional pump and treat How can the design and operation of a pump-
include hydraulic fracturing as well as chemical and-treat system be optimized and its perfor-
and biological enhancements. The pump-and-treat mance measured?
remediation approach is used at about three- When should variations and alternatives to
quarters of the Superfund sites where ground conventional pump-and-treat methods be
water is contaminated and at most sites where used?
cleanup is required by the Resource Conservation By presenting the basic concepts of pump-and-
and Recovery Act (RCRA) and state laws [Na- treat technology, this guide provides decision-
tional Research Council (NRC), 1994]. Although makers with a foundation for evaluating the
the effectiveness of pump-and-treat systems has appropriateness of conventional or innovative
been called into question (Sidebar 1), after two approaches. An in-depth understanding of
decades of use, this approach remains a necessary hydrogeology and ground-water engineering is
component of most ground-water remediation required, however, to design and operate a pump-
efforts and is appropriate for both restoration and and-treat system for ground-water remediation.
plume containment. Readers seeking more information on specific
This guide provides an introduction to pump- topics covered in this booklet should refer to the
and-treat ground-water remediation by addressing U.S. Environmental Protection Agency (EPA)
the following questions: documents listed at the end of this guide (Section
When is pump and treat an appropriate 9).
remediation approach?

Sidebar 1
Changing Expectations for the Pump-and-Treat Approach

Pump-and-treat systems for remediating ground water indicates that full restoration at some sites with
came into wide use in the early to mid-1980s. By the relatively simple characteristics is possible; moreover, at
early 1990s, evaluations by EPA (Keely, 1989; U.S. many sites, full restoration of ground-water quality can
EPA, 1989; Haley et al., 1991) and others (Freeze and be achieved for part of a site (NRC, 1994). For
Cherry, 1989; Mackay and Cherry, 1989) called into the example, Bartow and Davenport (1995), in a review of
question the performance of pump-and-treat systems. 37 applications of pump-and-treat systems in Santa
The general "failure" of the pump-and-treat approach Clara Valley, California, found that one site had
was identified as its inability to achieve "restoration" achieved maximum contaminant levels (MCLs) for all
(i.e., reduction of contaminants to levels required by contaminants and about one-third achieved, or were
health-based standards) in 5 to 10 years, as anticipated near, MCLs for one or more parameters. Bartow and
in the design phase of projects. Although a variety of Davenport's conclusion that pump-and-treat systems had
factors contributed to this shortcoming, tailing and significantly reduced the mass of volatile organic
rebound (Section 4) represented the major barrier to contaminants (VOCs) in the region's ground water
achieving remediation goals. Pump-and-treat systems indicates how expectations regarding the technology
were criticized more pointedly by Travis and Doty have changed.
(1990), who asserted as a "simple fact" that "contami-
nated aquifers cannot be restored through pumping and
Combining the pump-and-treat approach with in situ
bioremediation (see Section 7.4) provides further
opportunities for improving the effectiveness of ground-
Expectations for the effectiveness of pump-and-treat water cleanup. For example, Marquis (1995) suggested
technology, however, may have been too high. Ground- that in situ bioremediation used with the pump-and-treat
water scientists and engineers generally agree that approach should always be considered as an option for
complete aquifer restoration is an unrealistic goal for remediation of sand and gravel aquifers contaminated
many, if not most, contaminated sites. Nonetheless, with biodegradable organic compounds, especially
further experience with pump-and-treat systems volatile aromatic and polyaromatic hydrocarbons.

2. Appropriate Use of Pump-and-Treat Technology
Pump-and-treat systems are used primarily to ciently that the aquifer complies with cleanup
accomplish the following: standards or the treated water withdrawn from
Hydraulic containment. To control the the aquifer can be put to beneficial use.
movement of contaminated ground water, Although hydraulic containment and cleanup
preventing the continued expansion of the can represent separate goals, more typically,
contaminated zone. Figure 1 illustrates three remediation efforts are undertaken to achieve a
major configurations for accomplishing combination of both. For example, if restoration is
hydraulic containment: (1) a pumping well not feasible, the primary objective might be
alone, (2) a subsurface drain combined with a containment. In contrast, where a contaminated
pump well, and (3) a well within a barrier well is used for drinking water but the contami-
wall system. nant source has not been identified, treatment at
Treatment. To reduce the dissolved contami- the wellhead might allow continued use of the
nant concentrations in ground water suffi- water even though the aquifer remains contami-




Static Water Table

Plume Zone Limit






Static Water Table


Plume Drain Capture

Zone Limit






Upgradient Downgradient
Barrier Wall Barrier Wall

Figure 1. Plume Barrier

Examples of hydraulic Wall Capture
Zone Limit
containment in a plan 99


view and cross section

using pump-and-treat
technology: (a) pump well, Plume

(b) drain, and (c) well

within a barrier wall
system (after Cohen et
al., 1994)

3. Smart Pump-and-Treat Techniques
A fundamental component of any ground-water material to the ground water. Vapors also might
remediation effort using the pump-and-treat migrate to the water table and contaminate ground
approach is contaminant removal or control. Thus, water without infiltration.
effective remediation of ground water using Source removal is the most effective way to
pump-and-treat technology requires knowledge of prevent further contamination. Where inorganic or
contaminants and site characteristics. Addition- organic contaminants are confined to the vadose
ally, the remediation plan should call for imple- zone, removal is usually the preferred option.
mentation of dynamic system management based When removal is not feasible, as is often the case
on a statement of realistic objectives (Hoffman, with dense non-aqueous phase liquids (DNAPLs)
1993). residing below the water table, containment is an
3.1. Contaminant Removal/Control essential initial step in remediation. In some
Any ground-water cleanup effort will be situations, containment can be achieved through
undermined unless inorganic and organic con- capping, which prevents or reduces infiltration of
taminant sources are identified, located, and rainfall through the contaminated soil. Capping
eliminated, or at least controlled, to prevent can be ineffective if water table fluctuations occur
further contamination of the aquifer. Toxic within the zone of contamination or when NAPL
inorganic substances may serve as a continuing vapors are present.
source of contamination through mechanisms 3.2. Thorough Site Characterization
such as dissolution and desorption. At many Comprehensive characterization of the contami-
contaminated sites, organic liquids are a major nated site serves two major functions:
contributor to ground-water contamination.
Figure 2 illustrates four common types of con- Accurately assessing the types, extent, and
taminant plumes, each characterized by the forms of contamination in the subsurface
liquids density relative to water and the degree to increases the likelihood of achieving treat-
which the liquid mixes with water. Even when the ment goals. This requires an understanding of
organic liquid resides exclusively in the vadose the physical phases in which contaminants
zone (i.e., the area between the ground surface exist (mainly sorbed and aqueous phases for
and the water table) it can serve as a source of inorganic contaminants, and sorbed, NAPL,
ground-water contamination. In such situations, aqueous, and gaseous phases for organic
contamination occurs when percolating water liquids) and quantification of the distribution
comes in contact with the liquid (sometimes between the phases. Indeed, inadequate site
called product) or its vapors and carries dissolved characterization has undermined some pump-


Figure 2.
Contaminant plumes
as a function of
density and
miscibility with (b)
ground water:
(a) light liquids
(gasoline and
methanol) create
contaminant plumes
that tend to flow in
the upper portions of
an aquifer; (b) dense
liquids (perchloro-
ethylene [PCE] and
ethylene glycol)
create a plume that
contaminates the full
thickness of an
aquifer (adapted
from Gorelick et al.,

and-treat efforts; for instance, when after a greatly by performing two- and three-dimensional
few years greater quantities of contaminants computer modeling of the subsurface. Figure 6
had been removed than were identified in the shows a conceptual model of a site developed by
initial site assessment. combining contour visualization of a contaminant
A thorough, three-dimensional characteriza- plume of benzene with subsurface lithologic logs
tion of subsurface soils and hydrogeology, (Sidebar 3). EPAs SITE3D software, being
including particle-size distribution, sorption developed by the National Risk Management
characteristics, and hydraulic conductivity, Laboratorys Subsurface Protection and Remedia-
provides a firm basis for appropriate place- tion Division, allows three-dimensional visualiza-
ment of pump-and-treat wells. Such informa- tion of contaminant plumes (Figure 7). Statistical
tion is also required for evaluating the extent software developed by EPA such as Geo-EAS and
to which tailing and rebound may present GEOPACK for geostatistical analysis and con-
problems at a site (Section 4). touring of ground-water contaminant data and
GRITS/STAT for analysis of contaminant concen-
Three-dimensional characterization techniques trations are among the many computer-based tools
include primarily indirect observations, using available for analyzing subsurface data.
surface and borehole geophysical instruments and
cone-penetration measurements, and direct 3.3. Dynamic Management of the
sampling of soil and ground water. Important Well Extraction Field
advances in soil sampling technology have been To be effective, pump-and-treat efforts must go
made relatively recently, such as continuous beyond initial site characterization, using infor-
samplers used with a hollow-stem auger (Figure mation gathered after remediation operations are
3) and smaller continuous-core, direct-push under way to manage the well extraction field
equipment that also can be used to collect ground- dynamically. For instance, information collected
water samples without installing wells (Figure 4). while drilling and installing extraction wells,
Vibratory drilling methods are another innovative operating pumping wells, and tracking changes in
technique for collecting soil cores and ground- water levels in monitoring wells (Section 6.4) and
water samples. Additionally, sensitive borehole contaminant concentrations in observation wells
flow meters that allow measurement of vertical can refine the portrayal of the site.
changes in hydraulic conductivity in a borehole Dynamic management of the well extraction
represent an important recent development. These field based on more comprehensive information
techniques allow subsurface mapping to be can provide both economic and environmental
generated with a level of detail that generally benefits. In general, additional information about
would be prohibitively expensive using conven- the site and the pump-and-treat effort allows
tional drilling and sampling methods. Figure 5 operators to make informed decisions about the
presents a conceptual diagram of trichloroethene efficient use of remediation resources. More
(TCE) contamination at a complex site specifically, this flexible site management ap-
(Sidebar 2) developed from extensive use of proach may facilitate greater success in hydraulic
direct-push sampling techniques. containment (Section 5). Ultimately, the time
Moreover, if sufficient data are obtained, the required to achieve cleanup goals might be
interpretation of subsurface data can be enhanced minimized.

Figure 3.
This hollow-stem
auger is fitted
with a 5-foot
sampling tube
that collects a
continuous core
as the auger
allowing detailed
and accurate
observation of
lithology. When
drilling is
completed, a
monitoring well
also can be

Figure 4. Hydraulic or vibratory direct-push rigs can be installed on vans, small trucks, all-terrain vehicles, or
trailers and allow collection of continuous soil cores and depth-specific ground-water samples for
detailed subsurface mapping if contaminants are generally confined to depths of less than 15 meters.
(Photo courtesy of Geoprobe Systems.)

A key component of the dynamic management Phasing the construction of extraction and
approach is the effective design and operation of monitoring wells so that information obtained
the pump-and-treat system. The following tech- from operation of the initial wells informs
niques can be useful in this regard: decisions about siting subsequent wells
Using capture zone analysis, optimization (Section 6.2).
modeling, and data obtained from monitoring Phasing pumping rates and the operation of
the effects of initial extraction wells to individual wells to enhance containment,
identify the best locations for wells (Section avoid stagnation zones (Section 5.2.3), and

Figure 5. Conceptual diagram of DNAPL (TCE) based on soil and ground-water sampling in a heterogeneous
sand and gravel aquifer. The extreme difficulty in cleaning up this site, which includes five distinct forms
of TCE (vapors and residual product in the vadose zone; pooled, residual, and dissolved product in the
ground water) led to modification of the pump-and-treat system for hydraulic containment rather than
restoration (adapted from Clausen and Solomon, 1994).

ensure removal of the most contaminated Sciences (NRC, 1994) has identified three major
ground water first (Section 6.3). classes of sites based on hydrogeology (Sidebar
3.4. Realistic Cleanup Goals 2) and contaminant chemistry (see Table 1):
Unrealistic expectations for the pump-and-treat Class A. Sites where full cleanup to health-
approach can lead to disappointments in system based standards should be feasible using
performance (Sidebar 1). Indeed, a cleanup goal current technology. Such sites include homo-
that is realistic for one site may not be reasonable geneous single- and multiple-layer aquifers
elsewhere. The Committee on Ground Water involving mobile, dissolved contaminants.
Cleanup Alternatives of the National Academy of

Sidebar 2
Major Types of Hydrogeologic Settings
The Committee on Ground Water Cleanup Alterna- Fractured aquifers typically consist of
tives of the National Academy of Sciences has defined low-permeability rock where most ground-water
three major hydrogeologic settings for evaluating the flow is in joints and fractures.
technical feasibility of ground-water cleanup based on Single-layered aquifers are less complex than
the degree of uniformity of the aquifer material and multiple-layered aquifers, which are separated by
layering (Table 1). less-permeable strata, because the possibility of cross
Homogeneous aquifers consist of materials that do contamination between aquifers by either upward or
not vary significantly in their water-transmitting downward movement becomes a consideration. Note
properties. Contaminant movement in homogeneous that the term aquifer is used in this guide in a broad
aquifers is largely a function of the hydraulic sense to include any area within the saturated zone
conductivity of the aquifer. For example, a homoge- where the presence of ground-water contamination is of
neous aquifer might comprise permeable, well-sorted sufficient concern to require remediation. Figure 2
sands or gravels. illustrates a homogenous, single-layer aquifer.
Heterogeneous aquifers consist of materials that vary Figure 15c (described in Section 4.2.5) illustrates a
in their water-transmitting properties laterally, two-layered homogenous aquifer. Figure 5 illustrates a
vertically, or in both directions. Contaminants in multiple-layered heterogeneous aquifer. The challenge
heterogeneous aquifers move preferentially in the of ground-water cleanup increases along with aquifer
high-permeability zones, resulting in more rapid complexity because of difficulties in delineating
transport than would be expected based on the contaminant sources and pathways and the increased
average hydraulic conductivity of the aquifer. A sand likelihood of tailing and rebound effects (Section 4).
aquifer with lenses of silts and clays is an example of
a heterogeneous aquifer.

Class B. Sites where the technical feasibility by free-product light nonaqueous phase
of complete cleanup is likely to be uncertain. liquids (LNAPL) or DNAPL and single- or
This class includes a wide range of hydrogeo- multiple-layered heterogeneous aquifers
logic settings and contaminant types that do contaminated by a free-product DNAPL
not fall into classes A or C. (Sidebar 4).
Class C. Sites where full cleanup of the Typically, preliminary ground-water cleanup
source areas to health-based standards is not efforts at contaminated sites are focused on
likely to be technically feasible. Such sites standards established for drinking water, such as
include fractured-rock aquifers contaminated federal or state maximum contaminant levels

Figure 6.
GEOS computer
screen showing
plume in relation
to subsurface
stratigraphy (see
text for

(MCLs) or nonzero MCL goals (MCLGs). EPA removal of as much of the DNAPL as is feasible,
has established procedures, however, by which containment of the remaining DNAPL, and
efforts can target alternative goals at Superfund treatment of the aqueous contaminant plume
and RCRA sites using alternate concentration outside the containment area. Consequently, even
limits (ACLs) where ground-water discharges into at Class C and Class B sites where restoration is
nearby surface water (U.S. EPA, 1988) or demon- not feasible, application of some form of the
strating the technical impracticality (TI) of pump-and-treat approach may be required either
ground-water cleanup (U.S. EPA, 1993; Feldman to help contain the contaminant source and
and Campbell, 1994). At DNAPL sites where the aqueous-phase plume or to clean up the contami-
TI of ground-water cleanup has been demon- nated ground water outside the containment area.
strated, the remedial strategy might call for

Sidebar 3
Computer Graphics as a Site Characterization Tool

Computer visualization can help focus attention on lies near the cross section (see lower center log) reveals
the types of additional information needed when that the aquifer is missing at this point. This suggests
characterizing a site before initiating a remediation the absence of a concentration gradient between the two
effort. For example, in Figure 6 the contoured benzene wells, indicating that two different sources may be
data, collected from the sand/gravel #1 aquifer (yellow involved.
in the cross section in the upper right) shows two areas Further analysis of the spatial distribution of the sand/
of high concentration (i.e., MW-5 and MW-7). Does this gravel #1 aquifer, including flow directions as indicated
represent a contaminant plume from a single source, or by potentiometric heads, and possibly additional
does it indicate two contaminant plumes from separate sampling for benzene would be required to determine if
sources? An examination of the cross section (upper one or two sources are contributing to the contamina-
right of center) indicates that sand/gravel #1 aquifer at tion.
MW-7 has the lower high concentration of benzene,
This particular example also cautions against relying
suggesting that the two plumes might be related.
exclusively on computer-generated interpolations,
The cross section also shows, however, that the which can suggest features that are not actually present
aquifer is quite thin between the two monitoring wells. (i.e., continuity of the aquifer between MW-5 and
Indeed, examination of monitoring well MP-10, which MW-7).

Figure 7. EPAs SITE3D software, under development at the Ada, Oklahoma, laboratory, helps visualize in three-
dimensions a TCE contaminant plume at a Superfund Site. Yellows and reds indicate zone with highest
concentrations of TCE in ground water.

Table 1. Categories of Sites for Technical Infeasibility Determinations (NRC, 1994)

Contaminant Chemistry
Mobile Sorbed,
Dissolved Dissolved Strongly Separate Separate
(degrades/ Mobile, (degrades/ Sorbed, Phase Phase
Hydrogeology volatilizes) Dissolved volatilizes) Dissolved LNAPL DNAPL
Homogeneous, A A B B B B
single layer (1) (1-2) (2) (2-3) (2-3) (3)
Homogeneous, A A B B B B
multiple layers (1) (1-2) (2) (2-3) (2-3) (3)
Heterogeneous, B B B B B C
single layer (2) (2) (3) (3) (3) (4)
Heterogeneous, B B B B B C
multiple layers (2) (2) (3) (3) (3) (4)
Fractured B B B B C C
(3) (3) (3) (3) (4) (4)

Note: Shaded boxes at the left end (group A) represent types of sites for which cleanup of the full site to health-based standards should be feasible with
current technology. Shaded boxes at the right end (group C) represent types of sites for which full cleanup of the source areas to health-based
standards will likely be technically infeasible. The unshaded boxes in the middle (group B) represent sites for which the technical feasibility of
complete cleanup is likely to be uncertain. The numerical ratings indicate the relative ease of cleanup, where 1 is easiest and 4 is most difficult.

Sidebar 4
The Effect of NAPL Phases on Ground-Water Contamination

The light (LNAPLs) and dense (DNAPLs) immiscible pore spaces (Figure 8b and 9). The amount of
nonaqueous phase liquids shown in Figure 2 pose the residual NAPL remaining in the subsurface depends
most difficult problems for ground-water cleanup on the subsurface material and the type of NAPL.
because of their complex interactions with water and Residual saturation in the unsaturated zone typically
solids in the subsurface. Figure 5 illustrates these ranges from 10 to 20 percent of the subsurface
complexities when trichloroethene (TCE), a DNAPL has volume and in the saturated zone generally ranges
moved through a heterogeneous, multiple-layered from 10 to 50 percent (Cohen and Mercer, 1993).
alluvial aquifer (at a site in Tennessee). Four distinct Figure 5 differentiates residual TCE above and
forms, or phases, of TCE are evident: below the water table.
The NAPL emits a vapor phase in the unsaturated Free product exists where most of the pore space is
zone that moves by diffusion. DNAPL vapors tend to filled by the NAPL. It accumulates wherever a
sink until they reach impermeable layers (Figure 5) or barrier prevents downward movement. LNAPLs,
the water table. Even if the NAPL does not reach the such as gasoline, tend to float on top of the water
ground water, contamination can occur by dissolution table, whereas DNAPLS tend to sink until they reach
of the vapors directly into the ground water or by an impermeable layer (Figure 5).
water percolating through the unsaturated zone. Dissolved NAPL forms the aqueous contaminant
Residual NAPL remains after the free product has plume that moves in the direction of ground-water
moved through the subsurface by gravity or been flow. The residual NAPL and free product can serve
displaced by water (Figure 8a and b). Residual NAPL as a source of ground-water contamination as long as
exists as single- to complex-shaped blobs that fill they remain in the subsurface.

a b

Figure 8. Dark NAPL (Soltrol) and water in a homogenous micromodel after (a) the displacement of water by
NAPL and then (b) the displacement of NAPL by water, with NAPL at residual saturation (Wilson et al.,


Figure 9.
Photomicrographs of (a) a
single blob occupying one
pore body, and (b) a doublet
blob occupying two pore
bodies and a pore throat
(Wilson et al., 1990). b

4. Anticipating Tailing and Rebound Problems
The phenomena of tailing and rebound are Rebound is most problematic when a pump-
commonly observed at pump-and-treat sites. and-treat system attains the cleanup standard, but
Tailing refers to the progressively slower rate of concentrations subsequently increase to a level
decline in dissolved contaminant concentration that exceeds the standard.
with continued operation of a pump-and-treat 4.2. Contributing Factors
system (Figure 10). Rebound is the fairly rapid
The degree to which tailing and rebound
increase in contaminant concentration that can
complicate remediation efforts at a site is a
occur after pumping has been discontinued. This function of the physical and chemical characteris-
increase may be followed by stabilization of the
tics of the contaminant being treated, the subsur-
contaminant concentration at a somewhat lower
face solids, and the ground water. Major factors
and processes that contribute to tailing and
4.1. Effects of Tailing and Rebound rebound are discussed below.
on Remediation Efforts 4.2.1. Non-Aqueous Phase Liquids
Tailing presents two main difficulties for Although immiscible LNAPLs and DNAPLs
ground-water restoration: tend to be relatively insoluble in water, unfortu-
Longer treatment times. Without tailing, nately they often are sufficiently soluble to cause
contaminants theoretically could be removed concentrations in ground water to exceed MCLs.
by pumping a volume of water equivalent to Consequently, residual and pooled free-product
the volume of the contaminant plume (Figure NAPL will continue to contaminate ground water
10). The tailing effect, however, significantly that makes sufficient contact to dissolve small
increases the time pump-and-treat systems amounts from the NAPL surface (Figure 11a).
must be operated to achieve ground-water When ground water is moving slowly, contami-
restoration goals. Indeed, pumping may need nant concentrations can approach the solubility
to be conducted for hundreds of years rather limit for the NAPL (Figure 11c). Although pump-
than tens of years. and-treat systems increase ground-water velocity,
Residual concentrations in excess of the causing an initial decrease in concentration, the
cleanup standard. When tailing occurs, often decline in concentration will later tail off until the
initially the decline in the rate of contaminant NAPLs rate of dissolution is in equilibrium with
concentrations is fairly rapid, followed by a the velocity of the pumped ground water. If
more gradual decline that eventually stabilizes pumping stops, the ground-water velocity slows
at an apparent residual concentration level and concentrations can rebound, rapidly at first
above the cleanup standard (Figure 10). and then gradually reaching the equilibrium

Pumping On Pumping Off

Theoretical Removal
Without Tailing

Relative Concentration
Removal With Tailing

Apparent Residual


Figure 10. Concentration versus pumping duration or volume showing tailing and rebound effects (Cohen et al.,

concentration (Figure 11c), unless pumping is In heterogeneous aquifers, localized lenses of

resumed. low-permeability strata may cause pools of
As shown in Table 1, DNAPL contamination in free product to develop throughout the
heterogeneous and fractured aquifers is the most saturated zone (Figure 12). Low-permeability
intractable. The reasons for this are strata also may cause extensive lateral move-
ment of the DNAPL. DNAPL pools are
DNAPLs create an unstable wetting front in especially problematic because the contami-
the subsurface, with fingers of more rapid nant will dissolve even more slowly than
vertical flow speeding the movement deeper residual DNAPL. It may take tens of years to
into the saturated zone (Figure 12). (This also remove 1 cm of contaminant from a DNAPL
makes accurate delineation of zones of pool (NRC, 1994).
residual contamination extremely difficult in
homogeneous aquifers.) 4.2.2. Contaminant Desorption
The movement of many organic and inorganic
If the volume of DNAPL exceeds the residual
contaminants in ground water is retarded by
saturation capacity of the unsaturated and
sorption processes that cause some of the dis-
saturated zones, the DNAPL will reach lower
solved contaminant to attach to solid surfaces.
permeability materials and form pools of free
The amount of contaminant sorbed is a function
product (Figure 5).
of concentration, with sorption increasing as
concentrations increase, and the sorption capacity


Contaminant Free Product

Solid Grain

Liquid-Liquid Partitioning



Organic Carbon or Desorption of

Mineral Oxide Surface Adsorbed Contaminants


Figure 11. Equilibrium Concentration

Contaminants are mobilized
Dissolved Contaminant Concentration

when ground water that is

undersaturated with a
Low Ground-Water Velocities and Long Contact
contaminant comes in Times Produce High Contaminant Concentrations
contact with a NAPL (a) or (approaching equilibrium) in Ground Water
contaminant sorbed on an
organic carbon or mineral
surface (b). High ground-
water velocities and short High Ground-Water Velocities and Short Contact Times
contact times will result in Produce Low Contaminant Concentrations in Ground Water
low contaminant
concentrations, and low
velocities and long contact
Contact Time Increases to Right
times will result in high
contaminant concentrations Ground-Water Velocity Increases to Left
(c) (adapted from Gorelick et
al., 1993). (c)

Figure 12. Laboratory model of the transport of DNAPL contaminant through an aquifer with varying permeability;
note the concentration of downward movement in fingers and the DNAPL pools above the low-
permeability zones (the horizontal discs). (Source: U.S. EPA National Risk Management Research

of the subsurface materials. Sorbed contaminants concentrations during pumping as well as rebound
tend to concentrate on organic matter and clay- after pumping stops.
sized mineral oxide surfaces (Figure 11b). Sorp- 4.2.3. Precipitate Dissolution
tion is a reversible process, however. Thus, as As with sorption-desorption reactions, precipi-
dissolved contaminant concentrations are reduced tation-dissolution reactions are reversible. Thus,
by pump-and-treat system operation, contami- large quantities of inorganic contaminants, such
nants sorbed to subsurface media can desorb from as chromate in BaCrO4, may be found with
the matrix into ground water. Contaminant crystalline or amorphous precipitates in the
concentrations resulting from sorption and subsurface (Palmer and Fish, 1992). Figure 13
desorption show a relationship to ground-water illustrates a tailing curve where the contaminant
velocity and contact time similar to that of NAPLs concentration is controlled by solubility. In this
(Figure 11c), causing the tailing of contaminant situation, if pumping stops before the solid phase
is depleted, rebound can occur.


Relative Dissolved Concentration

Contaminant Concentration
Controlled by Solubility

Reserve Depleted


Pumping Duration or Volume Pumped

Figure 13. Dissolved contaminant concentration in ground water pumped from a recovery well versus time in a
formation that contains a solid-phase contaminant precipitate (Palmer and Fish, 1992).

4.2.4. Matrix Diffusion that the time required to reduce the concentration
As contaminants advance through relatively of TCE to 10 percent of the initial concentration
permeable pathways in heterogeneous media, would be 6 years for a clay lens 1 foot thick, 25
concentration gradients cause diffusion of con- years for a clay lens 2 feet thick, and 100 years
taminant mass into the less permeable media for a clay lens 4 feet thick. The significance of
(Gillham et al., 1984). Matrix diffusion is most matrix diffusion increases as the length of time
likely to occur with dissolved contaminants that between contamination and cleanup increases. In
are not strongly sorbed, such as inorganic anions heterogeneous aquifers, matrix diffusion contribu-
and some organic chemicals. During a pump-and- tions to tailing and rebound can be expected, as
treat operation, dissolved contaminant concentra- long as contaminants have been diffusing into
tions in the relatively permeable zones are re- less-permeable materials.
duced by advective flushing, causing a reversal in 4.2.5. Ground-Water Velocity Variation
the initial concentration gradient and slow diffu- Tailing and rebound also result from the vari-
sion of contaminants from the low to high perme- able travel times associated with different flow
ability media. Figure 14, based on theoretical paths taken by contaminants to an extraction well
calculations of TCE concentrations in clay lenses (Figure 15a-c). Ground water at the edge of a
of varying thickness, shows that diffusion is a capture zone created by a pumping well travels a
slow process. For example, the figure indicates greater distance under a lower hydraulic gradient
Average Relative Concentration


0.6 Clay Lens Thickness = 1.2 Meters (4 ft)

0.6 Meter (2 ft)
0.2 0.3
(1 ft)
0 20 40 60 80 100
Time (yr)

Figure 14. Changes in average relative trichloroethene (TCE) concentrations in clay lenses of varying thickness
as a function of time (NRC, 1994).




Fast Moderate


Relative Dissolved Concentration

1.0 Tailing Due to
Different Travel
0.8 Times Along Flow
Paths to
Recovery Well
Figure 15. 0.6
Tailing resulting from
ground-water velocity
variations: (a) horizontal 0.4
variations in the velocity of
ground water moving 0.2
toward a pumping well
(Keely, 1989) lead to (b) 0
tailing as higher Pumping Duration or Volume Pumped
concentrations of ground
water in slower pathlines
mix with lower
Stratified Sand-Gravel Aquifer
concentrations in faster (c)
pathlines (Palmer and Fish,
1992); (c) in a stratified
sand and gravel aquifer, t0
tailing occurs at t1 when
clean water from the upper
gravel strata mixes with still-
contaminated ground water
in the lower sand strata
(Cohen et al., 1994). t0 t1

than ground water closer to the center of the computer codes can be used to assess the potential
capture zone (Figure 15a). Additionally, contami- for tailing and rebound effects from precipitation-
nant-to-well travel time varies as a function of the dissolution reactions.
hydraulic conductivity in heterogeneous aquifers Assessing the potential for removal or contain-
(Figure 15c). ment of free product may be the first priority at
4.3. Assessing the Significance of NAPL-contaminated sites, followed by assess-
Tailing and Rebound at a Site ment of the extent of residual NAPL contamina-
Determining realistic objectives for a pump- tion. For DNAPLs, residual saturation may extend
and-treat system requires sufficient site character- throughout the unsaturated and saturated zones
ization to define the complexity of the hydrogeo- (Figure 5). Typically, for LNAPLs most residual
logic setting (Sidebar 2) and the subsurface contamination is located in the vadose zone, but it
distribution of contaminants. Such information may also extend to the depth of the seasonal low
makes it possible for the system operator to assess water table. As Figure 16 shows, pumping to
whether conditions at the site will result in tailing remove free LNAPL product can cause residual
and rebound and to evaluate the extent to which NAPL to move deeper into the saturated zone.
these conditions are likely to increase the time Consequently, when removing free-product
needed to attain health-based cleanup standards. LNAPL that is floating on the water table, steps
The sorption characteristics of contaminants can should be taken to avoid or minimize movement
be assessed using batch sorption tests with aquifer of residual NAPL deeper into the saturated zone.
materials (Roy et al., 1992), although aquifer Berglund and Cvetkovic (1995) evaluated the
heterogeneity increases the difficulty of interpret- relative importance of the degree of heterogeneity
ing test results. For organics, the potential effects in hydraulic conductivity and mass transfer
of sorption can be assessed based on a literature processes and concluded that the rate of mass
review of contaminant properties and on site- transfer and the extent to which contaminants are
specific data on organic carbon in aquifer materi- sorbed on aquifer solids are the most important
als (Piwoni and Keely, 1990). Geochemical parameters that affect predicted cleanup time.

Figure 16. Zone of residuals created in former cone of depression after cessation of LNAPL recovery system
(Gorelick et al., 1993).

5. Effective Hydraulic Containment
Hydraulic containment is a design objective of 5.1. Ground-Water Barriers and Flow
nearly all pump-and-treat systems. Where restora- Control
tion of an aquifer to health-based standards is the Hydraulic containment can be accomplished by
overall objective, the primary goal of containment controlling the direction of ground-water flow
must be to prevent farther spread of the contami- with capture zones (Section 5.1.1) or pressure
nant plume during restoration efforts. Where ridges (Section 5.1.2) or by using physical
NAPLs are present, containment using hydraulic barriers (Section 5.1.3). Figure 17 illustrates a
and physical barriers might be the primary pump-and-treat system that uses all three types of
objective for cleanup efforts in the portion of the hydraulic controls: (1) the contaminant source
aquifer contaminated by free product and residual area is surrounded by a barrier wall, (2) extraction
NAPL (Figures 1c and 17). In such situations a wells around the margins of the dissolved plume
conventional pump-and-treat system might be capture the contaminated ground water, and (3)
used to restore the dissolved contaminant plume treated ground water is reinjected to create a
(Figure 17). pressure ridge along the axis of the contaminant
Effective hydraulic containment using pumping plume. Note that the pressure ridge in Figure 17
wells requires the creation of horizontal and serves the function of increasing pore-volume
vertical capture zones that draw all contaminated exchange rates rather than functioning as a
ground water to the wells (Section 5.1.1) or other barrier. Barrier pressure ridge systems are created
hydraulic barriers (Sections 5.1.2 and 5.1.3). by placing injection wells along the perimeter of a
Failure to take aquifer anisotropy into account contaminant plume.
(Section 5.2.1) or limitations in the ability to 5.1.1. Horizontal and Vertical Capture
create sufficient drawdown to establish capture Zones
zones (Section 5.2.2) may allow contaminants to Pumping wells provide hydraulic containment
escape from these systems. Additionally, stagna- by creating a point of low hydraulic head to which
tion zones created by pumping operations or the nearby ground water flows. The portion of an
use of injection wells can reduce the effectiveness aquifer where flow directions are toward a
of cleanup efforts (Section 5.2.3). The monitoring pumping well is called a capture zone. In an
of both hydraulic heads (Section 6.4.1) and isotropic aquifer, where hydraulic conductivity is
ground-water quality (Section 6.4.2) can provide the same in all directions, ground-water flow is
early indications that contaminants are not being perpendicular to the hydraulic head contours, also
contained. called equipotential lines (Figure 18b).

Limit of Dissolved Plume
Source Area

Initial Ground-Water Flow Direction

Barrier Wall
Injection Well

Extraction Well

Figure 17. Plan view of a mixed containment-restoration strategy. A pump-and-treat system is used with barrier
walls to contain the ground-water contamination source areas (e.g., where NAPL or waste may be
present) and then collect and treat the dissolved contaminant plume (Cohen et al., 1994).

A pumping well creates a zone of influence The extent to which the aquifer is heteroge-
where the potentiometric surface has been modi- neous (Sidebar 2) or anisotropic (Section
fied (Figure 18c). The capture zone is the portion 5.2.1).
of the zone of influence where ground water flows Whether the aquifer is confined or uncon-
to the pumping well (Figure 18d). Figure 15a fined.
shows how a capture zone creates flow lines of
varying velocity. The size and shape of a capture The pumping rate and whether other pumping
zone depend on the interaction of numerous wells are operating.
factors, such as Whether the screened interval of the well
The hydraulic gradient and hydraulic conduc- fully or partially penetrates the aquifer.
tivity of the aquifer. When the screened portion of a pumping well
fully penetrates an aquifer (Figure 1b), a two-
dimensional analysis to delineate the horizontal

(a) (b)

Zone of Influence



Capture Zone

(c) (d)

Figure 18. In an isotropic aquifer, ground-water flow lines (b) are perpendicular to hydraulic head contours (a).
Pumping causes drawdowns and a new steady-state potentiometric surface within the wells zone of
influence (c). Following the modified hydraulic gradients, ground water within the shaded capture zone
flows to the pumping well (d). (Cohen et al., 1994, adapted from Gorelick et al., 1993).

capture zone is usually sufficient. When a pump- wells within a contaminant plume supply the
ing well only partially penetrates an aquifer, upgradient or downgradient injection wells used
however, vertical capture zone analysis also is to create a pressure ridge.
required to determine whether the capture zone 5.1.3. Physical Barriers
will contain a contaminant plume. Figure 19 Physical barriers are constructed of low-
shows a vertical capture zone for a partially permeability material and serve to keep fresh
penetrating well. If the contaminant plume ground water from entering a contaminated
extended to the base of the aquifer, some contami- aquifer zone. They also help prevent existing
nants would bypass the well, despite the presence areas of contaminant from moving into an area of
of apparent upward gradients. In stratified aniso- clean ground water or releasing additional con-
tropic media (Section 5.2.1), the vertical hydraulic taminants to a dissolved contaminant plume. Most
control exerted by a partially penetrating well will systems involving physical barriers also require
be further diminished. ground-water extraction to ensure containment by
5.1.2. Pressure Ridge Systems maintaining a hydraulic gradient toward the
Pressure ridge systems are produced by inject- contained area (see Figure 1c). The advantage of
ing uncontaminated water into the subsurface physical barriers is that the amount of ground
through a line of injection wells located water that must be extracted is greatly reduced
upgradient or downgradient of a contamination compared to the amount when using hydrody-
plume. The primary purpose of a pressure ridge is namic controls, as described in Sections 5.1.1 and
to increase the hydraulic gradient and hence the 5.1.2. Major types of barriers include
velocity of clean ground water moving into the Caps (or covers), which are made of low-
plume, thereby increasing flow to the recovery permeability material at the ground surface,
wells, which serves to wash the aquifer. can be constructed of native soils, clays,
Upgradient pressure ridges also serve to divert the synthetic membranes, soil cement, bitumi-
flow of uncontaminated ground water around the nous concrete, or asphalt.
plume, and downgradient pressure ridges prevent
further expansion of the contaminant plume. Slurry trench walls, excavated at the proper
Typically, treated ground water from extraction location and to the desired depth while

97 972


4 97


Vertical Capture

Figure 19. Cross section showing equipotential contours and the vertical capture zone associated with ground-
water withdrawal from a partially penetrating well in isotropic media (Cohen et al., 1994).

keeping the trench filled with a clay slurry, capture zone of a pumping well. In an aquifer that
keep the trench sidewalls from collapsing and is assumed to be isotropic, the general direction of
backfilling with soil bentonite, cement ground-water flow should be perpendicular to the
bentonite, or concrete mixtures. hydraulic gradient (Figure 20a). If fractures cause
Grout curtains are created by injecting hydraulic conductivity to be higher in a north-
stabilizing materials under pressure into the south rather than an east-west direction, however,
subsurface to fill voids, cracks, fissures, or the direction of ground-water flow will diverge
other openings in the subsurface. Grout also from the direction of the hydraulic gradient
can be mixed with soil using larger augers. (Figure 20b). In this example, siting a pumping
well based only on the hydraulic gradient (Figure
Sheet piling cutoff walls are constructed by 20a) would result in its failure to capture any
driving sheet materials, usually steel, through portion of a contaminant plume, except in the
unconsolidated materials with a pile driver or immediate vicinity of the well.
more specialized vibratory drivers.
A contaminant plume that does not follow the
Knox et al. (1984) provide further information hydraulic gradient may indicate that anisotropy is
on the design and construction of physical influencing the direction of ground-water flow.
ground-water barriers. Aquifer heterogeneities, such as buried stream
5.2. Hydraulic Containment: Other channels that have a different direction than the
Special Considerations hydraulic gradient, also may allow the direction
Certain site conditions can allow contaminants of contaminant travel to diverge from the hydrau-
to escape from a hydraulic containment system if lic gradient. Computer programs, such as EPAs
they are not characterized and anticipated. Well Head Protection Area (WHPA) code, can be
useful for evaluating the potential effects of
5.2.1. Effects of Anisotropy
anisotropy on well capture zones. Figure 21
In anisotropic aquifers, hydraulic conductivity
shows such a simulation for three pumping wells.
varies with direction. In flat-lying sedimentary
In this case, with a vertical to horizontal anisot-
aquifers, hydraulic conductivity is often higher in
ropy ratio of 10:1, the orientation of the capture
a horizontal than a vertical direction. In fractured
zones shifts from northwest-southeast (isotropic)
rock and foliated metamorphic rocks, such as
to east-west (anisotropic).
schist, the direction of maximum and minimum
permeability is usually aligned parallel and 5.2.2. Drawdown Limitations
perpendicular, respectively, to foliation or bedding Under some conditions creating and maintain-
plane fractures (Cohen et al., 1994). Where ing an inward hydraulic gradient for a contami-
sedimentary strata and foliated media are inclined nant plume is problematic. In such situations,
or dipping, significant horizontal anisotropy may injection wells may be required to create pressure
be an aquifer characteristic. In anisotropic media, ridges (Section 5.1.2) or physical barriers may
the flow of ground water, as well as contaminants need to be installed (Section 5.1.3). Site condi-
moving with ground water, is usually not perpen- tions that might indicate the need for such mea-
dicular to the hydraulic gradient. sures include (Cohen et al., 1994)
Figure 20 illustrates how horizontal anisotropy Limited saturated thickness of the aquifer
in fractured rock can change the location of the Relatively high initial hydraulic gradient

(a) Isotropic Aquifer (b) Anisotropic Aquifer

Zone of


KX 0 G
Zone 20 Di ene
of Contribution r
H ec ra
G ydr tion l

Gen nd-Wat
ra au

di lic of
G rec d-W rau en
en ti


er on
G nd die

al o
ro H nt
a ra

Dire r Flow
un yd

at lic



Pumping KY=5KX Pumping



Water- Well Well
Table KX



Figure 20. Effect of fracture anisotropy on the orientation of the zone of contribution (capture zone) to a pumping
well (Bradbury et al., 1991).

Sloping aquifer base stagnation zone can develop upgradient from an

Very high aquifer permeability injection well, however, and form in low-perme-
ability zones, regardless of hydraulic gradient.
Low aquifer permeability When multiple extraction or injection wells are
Where these conditions exist and hydraulic involved, a number of stagnation zones may
containment is planned, particular care should be develop (Figure 22b). Stagnation zones caused by
taken during site characterization and pilot tests to low hydraulic gradients can be identified by
assess drawdown limitations. measuring hydraulic gradients, tracer movement,
5.2.3. Stagnation Zones and ground-water flow rates using downhole
Stagnation zones develop in areas where pump- flowmeters and through modeling analysis.
and-treat operations create low hydraulic gradi- Stagnation zones within a contaminant plume can
ents and, consequently, low ground-water veloci- reduce the efficiency of a pump-and-treat system;
ties. The stagnation zone associated with a single thus, minimizing stagnation is an important
extraction well is likely to be located objective of capture zone analysis and optimiza-
downgradient from the well (Figure 22a). A tion modeling (Section 6.1).






0 (ft)
0 2100 4200 6300 8400 10500

(ft) Anisotropic



Figure 21.
Capture zone
simulation of 1800
three pumping
wells for an
isotropic aquifer
(a) and 0 (ft)
anisotropy ratio of 0 2100 4200 6300 8400 10500
10:1 (b) using the
EPA WHPA code. (b)







11 10 9 8 6



Pump Stagnation
Well 1 Point
2 12 3 4
13 11

12 10
98 6 5

Hydraulic Head
Contour Map Ground-Water Velocity
Contour Map

Extraction Well Extraction Well
536 87 6 5 4
538 538 3 1
535 536 535 6
537 7 8
7 8 6
Injection Well
Injection 6 45
Well 10 3
Hydraulic Head
Contour Map 30
15 20
543 3
4 8 6
542 5 10
541 6 7
538 7
537 538 540 537 8 Ground-Water Velocity Contour Map
536 1 6
539 535 0



535 3
Extraction Well Extraction Well

Figure 22. Examples of stagnation zones (shaded where ground-water velocity is less than 4 L/T): (a) single
pumping well and (b) four extraction wells with an injection well in the center (Cohen et al., 1994).

6. Pump-and-Treat System Design and Operation
The basic operating principle of a pump-and- cal equations (WHPA) and the innovative analytic
treat system calls for locating a well (or wells) element method (CZAEM) that allows capture
and then pumping at rates that cause all water in a zone and ground-water pathline analysis. The
contaminant plume to enter the well rather than numerical MODFLOW and MODPATH models
continue traveling through the subsurface. Table 2 developed by the U.S. Geological Survey are
lists types of data required for evaluating the commonly used to model more complex hydro-
feasibility of using the pump-and-treat approach geologic settings. Cohen et al. (1994) identify a
at a contaminated ground-water site and then number of computer codes of potential value for
designing an appropriate system. This section capture zone analysis. More detailed information
describes the key aspects of designing and about specific models and EPA guidance on the
operating a pump-and-treat system for optimal use of models are available in references on
performance. Ground-Water Modeling at the end of this guide
6.1. Capture Zone Analysis and (Section 9). Sidebar 5 summarizes the results of
computer modeling performed to evaluate the
Optimization Modeling effect of different hydrogeologic conditions on the
In recent years, numerous mathematical models effectiveness of different types of well patterns.
have been developed or applied to compute
capture zone, ground-water pathlines, and associ- In addition, optimization programming methods
ated travel times to extraction wells or drains. For are being used increasingly to improve pump-and-
relatively simple hydrogeologic settings (homoge- treat system design (Gorelick et al., 1993). As
neous isotropic aquifers), analytical equations applied to the design of pumping systems, optimi-
solved manually, using graphical techniques or zation involves defining an objective function,
computer codes based on analytical solutions, such as minimizing the sum of pumping rates
may be adequate. For more complex sites, nu- from a number of wells. A set of restrictions, or
merical computer models may be required. These constraints, specify various conditions, such as
models provide insight to flow patterns generated maximum pumping rates and minimum hydraulic
by alternative pump-and-treat approaches and to heads at individual wells, that must be satisfied by
the selection of monitoring points and frequency. the optimal solution alternative. Hydraulic
The WHPA model (Blandford and Huyakorn, containment of a contaminant plume usually
1991) and Capture Zone Analytic Element Model requires only linear optimization methods, but
(CZAEM) (Haitjema et al., 1994; Strack et al., when contaminant concentrations are specified as
1994) developed by EPA are examples of rela- constraints, nonlinear methods are often required
tively simple computer software based on analyti- (Rogers et al., 1995). At the Lawrence Livermore

Table 2. Data Requirements for Pump-and-Treat Systems (Adapted from U.S. EPA, 1991)

Data Description Purpose(s) Source(s)/Method(s)

Hydraulic conductivities and To determine feasibility of Pumping test, slug tests,
storativities of subsurface extracting ground water; laboratory permeability tests
materials applicability of pump-and-treat

Contaminant concentrations To determine seriousness of the Soil and water quality sampling
and areal extent problem; existence of NAPL; data
applicability and evaluate

Contaminant/soil properties To determine mobility Published literature, laboratory

(density, aqueous solubility, properties; applicability of tests
octanol-water/carbon pump-and-treat approach
partitioning coefficient, soil
organic carbon content,
sorption parameters)

Types, thicknesses, and extent To develop conceptual design; Hydrogeologic maps, surficial
of saturated and unsaturated applicability/considerations for geology maps/reports, boring
subsurface materials implementation logs, geophysics

Depth to aquifer/water table To select appropriate extraction Hydrogeologic maps,

system type; consideration for observation wells, boring logs,
implementation piezometers

Ground-water flow direction To determine proper well Water level data,

and vertical/horizontal gradients locations/spacing considerations potentiometric maps
for implementation

Seasonal changes in ground- To locate wells and screened Long-term water level
water elevation intervals; considerations for monitoring

NAPL density/viscosity/ To predict vertical distribution Literature, laboratory

solubility; residual saturation of of contamination; consideration measurements
vadose zone and saturated zone for implementation and
evaluating effectiveness

Ground-water/surface water To determine impacts of surface Seepage measurements,

connection water stream gaging

Precipitation/recharge To calculate water balance; NOAA reports, local weather

consideration for implementing bureaus; onsite measurements
and evaluating effectiveness

Locations, screen/open interval To determine Well inventory, pumpage

depths, and pumping rates of impacts/interference; records
wells influenced by site considerations for implementing
and evaluating effectiveness

Sidebar 5
Computer Modeling of Well Patterns Versus Hydrogeologic Conditions

Satkin and Bedient (1988) used the U.S. Geological The three-spot, doublet, and double-cell well
Survey MOC model to evaluate the effectiveness of patterns are effective under low hydraulic gradient
seven different well patterns (Figure 23) for restoring conditions. These well patterns minimize cleanup
contaminated ground water under eight generic time, volume of water circulated, and volume of
hydrogeologic conditions. The hydrogeologic settings water treated.
were defined as various combinations of three major The three-spot well pattern performed better than
factors: maximum drawdown (high > 10 ft; low < 5 ft), any of the other well patterns studied under a high
hydraulic gradient (high = 0.008; low = 0.0008), and hydraulic gradient, high drawdown, and either a low
longitudinal dispersivity (high = 30 ft; low = 10 ft). or high dispersivity.
Because the contaminants were assumed to not interact
None of the well patterns investigated was able to
with aquifer solids, tailing and rebound effects were not
contain and clean up the contaminated plume in a
a consideration in the study. Major conclusions of the
setting with high gradient, low drawdown, and high
computer simulations include the following:
Significant differences in cleanup time were observed
The centerline well pattern is effective in achieving
using various well locations for a given well pattern.
up to 99 percent contaminant reduction under both
low and high gradient conditions, but it may present
a water disposal problem.
The five-spot well pattern was the least effective of
the well patterns studied.


Single Doublet Centerline

..Pumping Well x ..Injection Well


x x
x x x x x

x x

5-Spot Double Triangle Double Cell

Figure 23.
Major types of
well patterns
(Satkin and
Bedient, 1988).

National Laboratory (LLNL) site, Rogers et al. off, others turned on, and pumping rates
(1995) applied an innovative nonlinear optimiza- varied to ensure that contaminant plumes are
tion approach, using artificial neural networks and remediated at the fastest rate possible. Figure
a genetic algorithm, to evaluate more than 4 24 illustrates stagnation zones that would
million pumping patterns for the projects 28 develop at the LLNL site if a fixed pumping
extraction and injection wells. The three top- well configuration were used. With this
ranked patterns required 8 to 13 wells, with approach, remediation at the site would take
projected costs estimated at $41 to $53 million about 100 years (Figure 25). Computer
over the 50-year project life. Using these pumping modeling of adaptive pumping indicates that
patterns was estimated to cost from one-third to this technique should make it possible to
one-quarter the cost of using all 28 wells at an reduce the time required for site cleanup to
estimated cost of $155 million. about 50 years (Figure 25). Further refine-
6.2. Efficient Pumping Operations ments in design might shorten the time even
Removal of contaminated ground water should further (Hoffman, 1993).
be a dynamic process that uses information on the Pulsed pumping, which has the potential to
response of the ground-water system to improve increase the ratio of contaminant mass
the efficiency of pumping operations (Section removed to ground-water volume where mass
3.3). Elements of efficient pumping operations transfer limitations restrict dissolved contami-
can include nant concentrations. Figure 26 illustrates the
Combined plume containment and source concept of pulsed pumping. During the
remediation, which can be achieved through resting phase of pulse pumping, contaminant
the design of the initial pumping flow field. concentrations increase due to diffusion,
For example, at the LLNL site a line of desorption, and dissolution in slower moving
extraction wells at the downgradient margins ground water (Figure 11). Once pumping is
of the plume were established to prevent the resumed, ground water with a higher concen-
movement of contaminants toward municipal tration of contaminants is removed, thus
water-supply wells, while other wells were increasing mass removal during pumping.
located in the source areas where the contami- Special care must be taken to ensure that the
nant concentrations were highest. This limited hydraulic containment objective is met during
the area requiring remediation and maximized pump rest periods. Bartow and Davenport
contaminant removal (Hoffman, 1993). (1995) have reported that about 19 percent of
the pump-and-treat systems in Santa Clara
Phased construction of extraction wells, Valley, California, use some form of pulsed
which allows data on the monitored response pumping. A recent study by Harvey et al.
of the aquifer to pumping operations to be (1994), however, on the effects of physical
used in siting subsequent wells. parameters (e.g., the mass transfer rate
Adaptive pumping, which involves designing coefficient) concluded that pulsed pumping
the well field such that extraction and injec- provides little if any advantage over continu-
tion can be varied to reduce zones of stagna- ous pumping at an average rate.
tion. Extraction wells can be periodically shut

Figure 24. Ground-water flow line in the vicinity of conceptual pumping centers at Lawrence Livermore National
Laboratory superimposed on an isoconcentration contour map and showing areas of potential
stagnation (Cohen et al., 1994, after Hoffman, 1993).


Highest Residual VOC Concentration

Fixed Well

Figure 25.
Effect of adaptive
pumping on Adaptive
cleanup time at Pumping
Lawrence 10
Superfund site
(Cohen et al.,
1994, after 0 5 10 15 20 25 30 35 40 45 50
Hoffman, 1993). Time (yr)

6.3. Treating Contaminated Ground Biological. Biological treatment methods use

Water microorganisms to degrade organic com-
Once extraction wells have brought contami- pounds and materials into inorganic products.
nated water to the surface, treatment is relatively The methods may be applicable for treatment
straightforward, provided that appropriate meth- of ground water contaminated by organic
ods have been selected and the capacity of the compounds if concentrations are low enough
treatment facility is adequate. Table 3 summarizes and the biological processes are not inhibited.
the applicability of various treatment technologies The best established biological treatment
to ground water contaminated by any of the major methods include (1) activated sludge systems,
categories of inorganic and organic contaminants. (2) a sequencing batch reactor, (3) powdered
U.S. EPA (1995) describes conventional technolo- activated carbon in activated sludge (bio-
gies that have evolved from industrial wastewater physical system), (4) rotating biological
treatment and that have been implemented at full contactors, and (5) an aerobic fluidized bed
scale for treatment of contaminated ground water. biological reactor.
These methods fall into two main categories:


t 1 t 2 t 3 t 4 t 5 t 6 t 7 t 8


Figure 26. Contamination
The pulsed
pumping concept
(Cohen et al., 1994, 0
after Keely, 1989). Time

Physical/Chemical. Physical, chemical, or a measuring hydraulic heads to determine if the

combination of physical and chemical meth- pump-and-treat system creates inward gradients
ods can be used to remove contaminants from that prevent ground-water flow and dissolved
ground water. The most commonly used contaminant migration across the containment
methods include (1) air stripping, (2) acti- zone boundary, and (2) ground-water quality
vated carbon, (3) ion exchange, (4) reverse monitoring to detect any contaminant movement
osmosis, (5) chemical precipitation of metals, or increase of contaminant mass across the
(6) chemical oxidation, (7) chemically containment zone boundary. Aquifer restoration
assisted clarification, (8) filtration, and (9) monitoring mainly involves measurement of
ultraviolet (UV) radiation oxidation. contaminant concentrations in pumping and
Various emerging and innovative treatment observation wells to determine the rate and
technologies, such as electrochemical separation effectiveness of mass removal. Cohen et al.
and wet air oxidation, are being tested. The EPA (1994) provide more detailed guidance on moni-
reference sources identified for ground-water toring the performance of pump-and-treat sys-
treatment methods at the end of this guide (Sec- tems.
tion 9) provide additional information on estab- 6.4.1. Hydraulic Head Monitoring for
lished and innovative treatment technologies. Containment
6.4. Monitoring Performance In general, the number of observation wells
An appropriately designed monitoring program needed for monitoring inward hydraulic gradients
is essential for measuring the effectiveness of a in a containment area increases with site complex-
pump-and-treat system in meeting hydraulic ity and with decreasing gradients along the
containment and aquifer restoration objectives. In containment perimeter. Strategies for adequately
general, containment monitoring involves (1) monitoring inward gradients and hydraulic
containment include (Cohen et al., 1994)

Measuring hydraulic heads in three dimen- Supplementing hydraulic head data with flow-
sions using nested piezometers for detecting path analysis using potentiometric maps or
vertical gradients. As shown in Figure 19, particle tracking computer codes. Figure 28
partially penetrating wells may not create an shows that ground water can flow between
adequate vertical capture zone. Where leaky and beyond recovery wells even though
confining layers separate aquifers, hydraulic hydraulic heads throughout the mapped
gradients should be toward the contaminated aquifer are higher than the pumping level.
zone. Figure 27 illustrates observations from Conducting an analysis to determine if
a nest of piezometers at the Chem-Dyne containment is threatened or lost when
Superfund site in Ohio. Water levels from the hydraulic head data do not indicate a clear
deep piezometer are consistently about a foot inward gradient. Rose diagrams can be
higher than in the intermediate and shallow prepared to display the variation over time of
piezometers, indicating an upward gradient. hydraulic gradient direction and magnitude
Monitoring water levels in observation wells based on data from at least three wells (Figure
intensively during system startup and equili- 29). Even when the time-averaged flow is
bration to determine an appropriate measure- toward the pump-and-treat system, contain-
ment frequency. This may involve using ment can be compromised if contaminant
pressure transducers and dataloggers to make escapes from the larger capture zone during
near-continuous head measurements for a few transient events or if a net component of
days or weeks, then switching sequentially to migration away from the pumping wells
daily, weekly, monthly, and possibly quarterly occurs over time.
monitoring. Data collected during each phase 6.4.2. Ground-Water Quality Monitoring
should provide the justification for any for Containment
subsequent decrease in monitoring frequency. Monitor well locations and completion depths
Making relatively frequent hydraulic head should be selected to provide a high probability of
measurements when the pumping rates or detecting containment system leaks in a timely
locations are modified, or when the system is manner. Consequently, monitor wells with
significantly perturbed in a manner that has relatively close spacing are usually located along
not been evaluated previously. Significant or near the potential downgradient containment
new perturbations can arise from, for ex- boundary. Ground-water quality sampling usually
ample, unusual recharge, flooding, drought, is performed less frequently than the measuring of
and new offsite well pumping. hydraulic head because contaminant movement is
Measuring hydraulic head as close to the a slower process. Because ground-water quality
same time as possible when monitoring monitoring is more expensive than hydraulic head
inward hydraulic gradients or a potentiomet- monitoring, designing a cost-effective monitoring
ric surface so that data are temporally consis- plan requires special care. Strategies that may
tent. This ensures that differences in ground- help reduce costs without compromising the
water elevation within a network represent integrity of the program include (Cohen et al.,
spatial rather than temporal variations. 1994)

Table 3. Applicability of Treatment Technologies to Contaminated Ground Water (U.S. EPA, 1991)


Membrane Separation*
Chemical Oxidation

Gravity Separation
Activated Carbon
Steam Stripping

Ion Exchange


Air Stripping





Heavy metals X X X X X X X X

Hexavalent chromium X X X X X X X X X X X

Arsenic X X X X X X X X X

Mercury X X X X X X X X X X

Cyanide X X X X X X X X X X X

Corrosives X X X X X X X X X X X X X X X

Volatile organics X X X X X X X X X

Ketones X X X X X X X X X X X X

Semivolatile organics X X X X X

Pesticides X X X X


Dioxins X X X X X

Oil and grease/floating X X X X X X X X


Applicable Potentially Applicable X Not Applicable

*Technology includes several processes; reverse osmosis and ultrafiltration among others.


Water-Level Elevation (ft)



564 564.24


Jan Feb Mar Apr May June July Aug Sept Oct Nov Dec

Water-Level Measurement in Deep Piezometer Time-Averaged Water Level in Deep Piezometer
Water-Level Measurement in Intermediate Piezometer Time-Averaged Water Level in Intermediate Piezometer
Water-Level Measurement in Shallow Piezometer Time-Averaged Water Level in Shallow Piezometer
Daily Average Water Level From Hourly Data Recorded
in Shallow Piezometer

Figure 27. Nested piezometer hydrograph for 1992 at the Chem-Dyne Superfund site (Cohen et al., 1994, after Papadopulos &
Associates, 1993).

2.0 Constant-Head
2.5 Pumping
Elevation in
Each Well Is

.0 360.0 ft

Figure 28.
Ground-water 2.5
flow between and 363.0
beyond the
extraction wells,
resulting even
though hydraulic 0
5 364.
heads throughout 363.
the mapped
aquifer are higher
than the pumping
level (Cohen et
al., 1994).

Sampling more frequently and performing well as the maximum plausible outward
more detailed chemical analyses in the early hydraulic gradients. Consider more frequent
phase of the monitoring program, and using sampling of more permeable strata in which
the information gained to optimize sampling migration might occur relatively quickly as
efficiency and reduce the spatial density and compared to the sampling frequency for less
temporal frequency of sampling in the later permeable media.
phases. Focusing chemical analyses on site contami-
Monitoring ground-water quality in perimeter nants of concern and indicator constituents
and near-perimeter leak detection wells more after performing detailed chemical analyses
frequently than in wells that are at a greater during the remedial investigation or the early
distance from the contaminant plume limit. phase of a monitoring program. Conduct
Specifying sampling frequency based on more detailed chemical analyses less fre-
potential containment failure migration rates quently or when justified based on the results
that factor in hydraulic conductivity and of the more limited analyses.
effective porosity of the different media as

6.4.3. Aquifer Restoration Monitoring chemicals that may indicate the occurrence of
Aquifer restoration monitoring consists of three other processes of interest, such as dissolved
main elements: oxygen, carbon dioxide, and biodegradation
Ground-water sampling from all extraction products. These sampling data are important
wells and selected observation wells within for making adjustments for efficient well
the contaminant plume to interpret cleanup operation (Section 6.2).
progress. Parameters analyzed should include Periodic sampling and chemical analysis of
(1) the chemicals of concern, (2) chemicals aquifer materials from representative loca-
that could affect the treatment system, such as tions in the contamination zone to measure
iron, which can precipitate and clog treatment removal of nondissolved contaminants.
units if ground water is aerated, and (3)







West East

Figure 29.
Example display of ground-water
flow directions and hydraulic
gradients determined between
three observation wells (Cohen et
al., 1994). South

Regular sampling and analysis of treatment The simplest indicator of progress in removing
system influent and effluent to assess (1) ground-water contaminants is a plot of the
treatment system performance, (2) change in cumulative mass removed from the aquifer as
influent chemistry that may influence treat- measured by influent concentrations to the
ment effectiveness, and (3) dissolved con- treatment system. Figure 31 shows the cumulative
taminant concentration trends. Figure 30 mass of VOC removal at the Chem-Dyne site.
shows influent and effluent VOC concentra- Approximately 27,000 pounds of VOCs have been
tions for the first 6 years of operation at the removed since the system became operational. As
Chem-Dyne Superfund site treatment plant. is apparent from both Figures 30 and 31, however,
Influent concentrations data showed a large the rate of removal slowed significantly in the
drop in the first year, and then a more gradual sixth year. Consequently, removal of the remain-
decline over the next 5 years due to tailing ing one-third of the in-place mass will take much
effects (Section 4). longer than 6 years.

12 400

Concentration, in g/L
10 300
Concentration, in mg/L

8 200

6 100

4 0
1991 1992

0 1988 1989 1990 1991 1992


Influent Effluent

Figure 30. Influent and effluent VOC concentrations (mg/L) at the Chem-Dyne treatment plant from 1987 to 1992
(Cohen et al., 1994, after Papadopulos & Associates, 1993).


Mass of Priority Pollutant VOCs, in lb






1987 1988 1989 1990 1991 1992

Figure 31. Cumulative mass of VOCs removed from the aquifer at the Chem-Dyne site from 1987 to 1992 (Cohen
et al., 1994, after Papadopulos & Associates, 1993).

6.5. Evaluating Restoration Success Stage 2. Operation of the remediation system,

and Closure during which contaminant concentrations
Ground-water restoration, as operationally decline.
defined, is achieved when a predefined cleanup Stage 3. Conclusion of treatment after con-
standard is attained and sustained. Figure 32 taminant concentrations have remained below
outlines procedures for determining the success the cleanup standard for a sufficient period of
and/or timeliness of closure of a pump-and-treat time based on expert knowledge of the
system. U.S. EPA (1992) defines six stages of ground-water system and data collected
remediation using water quality data from a single during pump-and-treat operations.
well (Figure 33): Stage 4. Post-termination monitoring of water
Stage 1. Site evaluation to determine the need levels and contaminant concentrations to
for and conditions of a remedial action; define determine when the ground-water flow
cleanup standard. system is reestablished.

Define Attainment Objectives and
Cleanup Standard

Develop Sampling and Analysis

Plan for Performance

Monitor System

No No
Continue/ Is the Cleanup Demonstration of
Modify Standard Reached? Technical
Treatment Impracticability

Terminate Remedial
Treatment Action

Allow System to Reach


Verify the Attainment of

Assess and Revise Cleanup Standard
Treatment Design as

No Is the Cleanup
Standard Maintained
Figure 32. Over Time?
Determining the
success and/or Yes
timeliness of
closure of a
pump-and-treat Monitor as
system (Cohen et Necessary
al., 1994).

1.2 Treatment

Measured Ground-Water Concentration

Treatment End Sampling
Declare Clean or
0.8 Contaminated
Start 6
0.6 1 2

4 5



Figure 33. Stages of remediation in relation to example contaminant concentrations in a well at a pump-and-treat
site (U.S. EPA, 1992).

Stage 5. Sampling to assess attainment of the

Cohen et al. (1994) and U.S. EPA (1992)
cleanup standard. If the treatment standard is
address in more detail the types of statistical
not met, the treatment design may need to be
techniques that are required to analyze short-term
assessed and revised (Figure 32).
and long-term trends in contaminant concentra-
Stage 6. Declaration that the aquifer is clean tions.
or still contaminated based on data collected
during Stage 5.

7. Variations and Alternatives to Conventional Pump-and-Treat Methods
Numerous variations and enhancements of reducing the requirements for surface treat-
pump-and-treat systems are possible. Major types ment of contaminated ground water (Section
include 7.4).
Using trenches or drains in combination with Notably few alternatives to pump-and-treat
or to replace vertical pumping wells (Section systems are without requirements for continuous
7.1). Where site conditions are favorable (i.e., energy input for pumping fluids (Section 7.4).
shallow contamination), trenches are a 7.1. Alternative Methods for Fluid
commonly used method for intercepting
contaminated ground water.
Delivery and Recovery
Conventional pump-and-treat systems usually
Using horizontal wells or trenches to replace involve extraction wellsand possibly injection
or complement vertical wells (Section 7.1). wellsplaced vertically in an aquifer. Alternative
Recent developments in directional drilling methods of delivery and recovery of contaminated
technology make the use of horizontal or ground water might enhance the performance of a
inclined wells an attractive alternative ap- pump-and-treat system, especially while interim
proach. measures are undertaken, by improving the
Inducing fractures in the subsurface to effectiveness of containment. These methods also
improve the yield of wells (Section 7.1). might augment the performance of a variety of
Although widely used by the petroleum remedial actions selected as possible long-term
industry, the use of induced fractures is remedies. Major alternatives include
considered an emerging technology in Interceptor Trenches. After vertical wells,
ground-water remediation with applications trenches are the most widely used method for
limited to contaminated ground water in low- controlling subsurface fluids and recovering
permeability materials. contaminants. They function similarly to
Implementing vadose zone source control and horizontal wells, but also can have a signifi-
remediation, often as a necessary adjunct to cant vertical component, which cuts across
ground-water cleanup (Section 7.2). and can allow access to the permeable layers
Making chemical enhancements, which can in interbedded sediments. For shallower
have the potential to accelerate aquifer applications, trenches can be installed at
remediation (Section 7.3). relatively low cost using conventional equip-
ment. Recent innovations combine trench
Making biological enhancements, which can excavation and well screen installation into a
present opportunities for eliminating or

single step for depths up to 20 feet (U.S. EPA, an estimated cost of $1 million. Instead, a con-
1994). Where depth is not a constraint, tinuous excavation and completion system was
interceptor trenches are generally superior to installed for less than $350,000 (U.S. EPA, 1994).
vertical wells. In such situations, they are EPAs Manual Alterative Methods for Fluid
especially effective in low-permeability Delivery and Recovery (U.S. EPA, 1994) provides
materials and heterogeneous aquifers. more detailed information on design consider-
Horizontal and Inclined Wells. Relatively ations and applications of these methods.
recent advances in directional drilling tech- 7.2. Vadose Zone Source Control
nology, which use specialized bits to curve Removal of contaminants from the vadose
bores in a controlled arc, have revolutionized (unsaturated) zone is an essential part of any
the field of well design. Directional drilling remedial action plan to clean up contaminated
methods can create wellbores with almost any ground water. Major methods include
trajectory. Wells that curve to a horizontal Capping to reduce infiltration of precipitation.
orientation are especially suited to environ-
mental applications (Figure 34). Excavation to remove contaminated soil for
ex situ treatment, which is most commonly
Induced Fractures. EPA research has shown used where contaminants have not penetrated
that petroleum engineering technology used deeply into the subsurface.
to induce fractures for increased productivity
of oil wells also can improve the performance Soil vapor extraction (SVE), which is used to
of environmental wells. Induced fractures are extract volatile organic contaminants by
used mainly where low-permeability aquifer flushing with air, and bioventing, a SVE
materials create problems for the recovery of system in which the addition of nutrients
contaminants. further enhances the biodegradation of
organic contaminants. Both techniques,
Table 4 rates the potential applications of considered innovative technologies a few
alternative methods for delivery or recovery of years ago, are widely used.
subsurface fluids in relation to (1) access, (2)
depth, (3) recovered phases, (4) geology, and (5) In situ thermal technologies to enhance the
availability. Figure 35 illustrates two ways in mobility of volatile and semivolatile organic
which horizontal wells or trenches can be used to contaminants; for example, steam-enhanced
intercept a contaminant plume. In many applica- extraction and radio frequency heating are
tions, deciding between use of a trench or a promising innovative technologies.
horizontal well hinges on economic rather than 7.3. Physical and Chemical En-
technical issues, with trenches generally being hancements
more cost effective at depths less than 20 feet and Physical and chemical enhancements to pump-
horizontal wells being generally more cost and-treat systems primarily function by enhancing
effective at depths greater than 20 feet. Cost the mobility of contaminants, thus increasing their
savings can be substantial compared to vertical recovery in ground water that has been pumped to
well systems. For example, initial remediation the surface for treatment. Some chemical en-
plans at a site in North Carolina called for 100 hancements transform contaminants in place in
vertical wells to recover a hydrocarbon plume at the subsurface to reduce toxicity.





Figure 34. Some applications of horizontal wells: (a) intersecting flat-lying layers, (b) intercepting plume elongated
by regional gradient, (c) intersecting vertical fractures, and (d) access beneath structures (U.S. EPA,

Table 4. Issues Affecting Application of Alternative Methods for Delivery or Recovery (U.S. EPA, 1994)

Issue Horizontal Well Induced Fracture Trench


Fragile structures Minimal surface disturbance Evaluate effects of surface Excavation expected to be
over target displacement infeasible

Poor access over Standoff required Possible with horizontal Excavation expected to be
target well infeasible


<6m 1 m minimum depth 1-2 m minimum depth Installation with common


6-20 m Cost of guidance system Excavation costs increase

increases at >6 m with depth

>20 m No depth limit within No depth limit within Specialized excavation

environmental applications environmental applications methods required

Recovered Phase


LNAPL Requires accurate drilling; Best with access to Widely used to ensure
best if water table fluctuations individual fractures capture; accommodates
are minor water table fluctuations

DNAPL Requires accurate drilling and Caution; steeply dipping Assuming mobile phase
site characterization fractures may cause present and accurately
downward movement located

Vapor Consider omitting gravel pack Best with access to Requires tight seal on top
to save costs individual fractures of trench


Normally Smearing of bore wall may Induced fractures may be Large discharge expected
consolidated clay reduce performance vertical and limited in size relative to alternatives

Swelling clay Smearing of bore wall may Relatively large, gently Large discharge expected
reduce performance dipping fractures expected relative to alternatives

Silty clay till Smearing of bore wall may Relatively large, gently Large discharge expected
reduce performance dipping fractures expected relative to alternatives

Stratified sediment Anisotropy may limit vertical Stratification may limit Good way to access many
or rock influence of well upward propagation and thin beds or horizontal
increase fracture size partings

Table 4. (Continued)

Issue Horizontal Well Induced Fracture Trench

Vertically fractured Orient well normal to Good where induced Orient trench
sediment or rock fractures when possible fractures cross-cut natural perpendicular to natural
fractures fractures when possible
sediment and rock)

Coarse gravel Possible problems with hole Permeability enhancement Stability a concern during
stability; penetrating cobbles may be unnecessary excavation

Thick sand May be difficult to access top Permeability enhancement Stability a concern during
and bottom of formation; hole may be unnecessary excavation
stability problems

Rock Feasible, but drilling costs Widely used in oil, gas, Excavation difficult but
more in rock than in sediment and water wells drilled in blasting possible to make
rock trench-like feature

Availability 10 to 20 companies with capabilities; Several companies offer service; Shallow trench (<6 m)
nationwide coverage but may require nationwide coverage with installation widely available
equipment mobilization equipment mobilization from local contractors; deep
trench will require mobilization

Current Experience 150 to 250 wells at 50 to 100 sites 200 to 400 fractures at 20 to 40 1,000+ trenches at many
(Approximate) sites hundreds of sites
Good application
Moderately good
Fair, with possible technical difficulties
Poor; not recommended using available methods


Horizontal Well


Capture Zone


Horizontal Well



Figure 35. Two approaches using trenches or horizontal wells to intercept contaminant plumes (U.S. EPA, 1994).

7.3.1. Physical Enhancements In situ chemical treatment, which involves
Air sparging, also known as in situ aeration, is reactive agents that oxidize or reduce con-
an approach that is similar to soil vapor extrac- taminants, converting them to nontoxic forms
tion except that air is injected into the saturated or immobilizing them to minimize contami-
zone rather than the vadose zone (Figure 36). Air nant migration. This innovative technology is
sparging systems can effectively remove a still in the early stages of development.
substantial amount of volatile aromatic and The EPA report Chemical Enhancements to
chlorinated hydrocarbons in a variety of geologic Pump-and-Treat Remediation (Palmer and Fish,
settings, but significant questions remain about 1992) provides additional information on techni-
the ability of this technology to achieve health- cal issues related to this topic.
based standards throughout the saturated zone
(NRC, 1994). Thermal enhancements, such as 7.4. Biological Enhancements
steam and hot-water flooding, increase the Biological enhancements to pump-and-treat
mobility of volatile and semivolatile contami- systems stimulate subsurface microorganisms,
nants. Use of induced fractures (Section 7.1) is primarily bacteria, to degrade contaminants to
another form of physical enhancement to pump- harmless mineral end products, such as carbon
and-treat systems. dioxide and water. In situ bioremediation of
certain types of hydrocarbons (primarily petro-
7.3.2. Chemical Enhancements leum products and derivatives), encouraged by
Chemically enhanced pump-and-treats systems addition of oxygen and nutrients to the ground
require use of injection wells to deliver reactive water, is an established technology. Other readily
agents to the contaminant plume and extraction biodegradable substances, such as phenol, cresols,
wells to remove reactive agents and contaminants acetone, and cellulosic wastes, are also amenable
(Figure 37). The major types of chemical en- to aerobic in situ bioremediation. Key elements in
hancements are such a system are delivery of oxygen and nutri-
Soil flushing, which enhances recovery of ents by use of an injection well (Figure 38a) or an
contaminants with low water solubility, free- infiltration gallery (Figure 38b). A limitation of in
product and residual NAPLs, and sorbed situ bioremediation is that minimum contaminant
contaminants. Two major types of chemical concentrations required to maintain microbial
agents can be used: (1) cosolvents, which, populations may exceed health-based cleanup
when mixed with water, increase the solubil- standards, particularly where heavier hydrocar-
ity of some organic compounds, and (2) bons are involved.
surfactants, which may cause contaminants In situ bioremediation of chlorinated solvents is
to desorb and may increase NAPL mobility less well demonstrated because metabolic pro-
by lowering the interfacial tension between cesses for their degradation are more complex
the NAPL and water, increasing the solubil- than those for hydrocarbon degradation (NRC,
ity. Soil flushing is one of the most promis- 1994). Nonetheless, methanotrophs are able to
ing innovative technologies for dealing with degrade some chlorinated solvents under aerobic
separate phase DNAPLs in the subsurface conditions if methane is supplied as an energy
(NRC, 1994). source. Also, the ability of anaerobic bacteria to
degrade a variety of chlorinated solvents is well

Vent to
(a) Atmosphere

Air Compressor Vacuum

Vapor Extraction Vapor
Wells Treatment

Surface Soil/Cap

Unsaturated Zone

Saturated Zone
Contaminated Zone

Direction of Ground-
Streamtubes of Air Water Flow

Injection Point for Flushing Gas
Extraction of Contaminated Gas

Surface Soil/Cap

Unsaturated Zone

Figure 36.
Process diagram
for air sparging Contaminated Zone
with (a) vertical
wells, and (b) Saturated Zone
horizontal wells
(after NRC,


Recovery of
Reactive Agent

Injection Treatment
of Reactive Extraction of
Agents Reactive Agents
and Contaminants

Figure 37.
Schematic of chemical
enhancement of a pump-and-
treat system. Key areas of
concern are shown in boxes. Delivery Reaction Removal
In some cases, the reactive
agent will be recovered and
reused (Palmer and Fish,

documented. Two major obstacles to the use of 7.5. Alternatives to the Pump-and-
anaerobic processes for in situ bioremediation are Treat Approach
that (1) hazardous intermediate degradation Nearly all approaches to ground-water cleanup
products can accumulate, and (2) undesirable involve some degree of ground-water pumping.
water quality changes, such as dissolution of iron Even when containment is the primary objective,
and manganese, can occur. low-flow pump-and-treat systems are usually
EPA reference sources identified at the end of required to prevent the escape of contaminated
this guide (Section 9) that are particularly relevant water from the confined area. Two remediation
to in situ bioremediation include Norris et al. approaches that eliminate pumping as a compo-
(1993), Sims et al. (1992), and U.S. EPA (1993, nent of the system are (1) intrinsic bioremedia-
1994). tion, and (2) in situ reactive barriers. Although
both of these methods show promise, they are still


Nutrient Addition Tank

To Sewer or

Air Compressor
Course Sand

Production Well Injection Well

Water Table

Spilled Materials

Air Compressor or
Hydrogen Peroxide
Tank Nutrient Addition

Infiltration Gallery

Figure 38. Trapped Hydrocarbons

Two types of Table
aerobic in situ
systems: (a)
injection well with Recirculated Water
sparger, (b) and Nutrients
infiltration gallery
(Sims et al., Recovery Well
1992, after Monitoring Well
Thomas and
Ward, 1989).

in development and their effectiveness remains to 7.5.2. In Situ Reactive Barriers
be demonstrated. The concept of using permeable in situ reactive
7.5.1. Intrinsic Bioremediation barriers to treat a contaminant plume as it moves
Intrinsic bioremediation relies on indigenous through an aquifer under natural hydraulic
microbes to biodegrade organic contaminants, gradients (Figure 39c and 39d) was first suggested
without human intervention in the form of supply- by McMurty and Elton (1985), but it has only
ing electron acceptors, nutrients, and other recently begun to receive significant attention
materials. The processes that occur are the same from the research community (Starr and Cherry,
as those in engineered bioremediation systems, 1994). The funnel-and-gate concept, which
but they occur more slowly. A decision to refrain combines impermeable barriers to contain and
from active site manipulation does not eliminate channel the flow of the contaminant plume toward
the need to conduct ground-water sampling within the reactive barrier has received the most attention
the contaminant plume to document that biodeg- because numerous possible configurations can be
radation is occurring. Moreover, sampling would developed to address different types of contami-
still need to be performed outside the contami- nant plumes and geologic settings (Figure 40).
nated area to identify any offsite migration of Depending on the contaminants present in the
contaminants that might require initiation of more plume, the reactive zone uses a combination of
active remedial measures (Figure 39b). There is a physical, chemical, and biological processes.
greater risk of failure with intrinsic bioremedia- The great promise of in situ reactive barriers is
tion compared to engineered bioremediation that they will require little or no energy input once
because no active measures are used to control the installed, yet provide more active control and
contaminant plume. The possible perception that treatment of the contaminant plume than intrinsic
intrinsic bioremediation is the equivalent to doing bioremediation. The main engineering challenges
nothing is also a barrier to its acceptance (NRC, involve provision of suitable amounts of reactive
1994). materials in a permeable medium and proper
placement to avoid short-circuiting the contact
between the gate and the cutoff wall.

Contaminant Source
(a) Zone
Extraction Well


Ground-Water Plume

(c) In Situ Reaction Curtain'


(d) In Situ Reactor


Cutoff Wall Remediated

'Funnel' Plume

Figure 39. Alternative ground-water plume management options: (a) pump-and-treat system, (b) intrinsic
bioremediation, (c) in situ reaction curtain, (d) funnel-and-gate system (adapted from Starr and
Cherry, 1994).

(a) Single Gate System

Multiple Gate System Multiple Reactor Systems

Fully Penetrating Gate Hanging Gate

Figure 40. Funnel-and-gate configurations (Starr and Cherry, 1994).

8. References
Bartow, G. and C. Davenport. 1995. Pump-and- Cohen, R.M., A.H. Vincent, J.W. Mercer, C.R.
Treat Accomplishments: A Review of the Faust, and C.P. Spalding. 1994. Methods for
Effectiveness of Ground Water Remediation Monitoring Pump-and-Treat Performance.
in Santa Clara Valley, California. Ground EPA/600/R-94/123. R.S. Kerr Environmental
Water Monitoring and Remediation Research Laboratory, Ada, OK. 102 pp.
Clausen, J.L. and D.A. Solomon. 1994. Character-
Berglund, S. and V. Cvetkovic. 1995. Pump-and- ization of Ground Water Plumes and DNAPL
Treat Remediation of Heterogeneous Aqui- Sources Using a Driven Discreet-Depth
fers: Effects of Rate-Limited Mass Transfer. Sampling System. Ground Water Manage-
Ground Water 33(4):675-685. ment 18:435-445 (Proc. of 8th Nat. Outdoor
Action Conf. on Aquifer Remediation,
Blandford, T.N. and P.S. Huyakorn. 1991. WHPA: Ground Water Monitoring and Geophysical
Modular Semi-Analytical Model for the Methods).
Delineation of Wellhead Protection Areas,
Version 2.0. Office of Ground Water Protec- Feldman, P.R. and D.J. Campbell. 1994. Evaluat-
tion; Available from EPA Center for Subsur- ing the Technical Impracticality of Ground-
face Modeling Support, Ada, OK. Version 1.0 Water Cleanup. Ground Water Management
was released in 1990 [Four modules: 18:595-608 (Proc. of 8th Nat. Outdoor Action
MWCAP, RESSQC, GPTRAC, MONTEC; Conf. on Aquifer Remediation, Ground Water
most current disk version is 2.1] Monitoring and Geophysical Methods).
Bradbury, K.R., M.A. Muldoon, A. Zaporozec, Freeze, R.A. and J.A. Cherry. 1989. What Has
and J. Levy. 1991. Delineation of Wellhead Gone Wrong? Ground Water 27(4):458-464.
Protection Areas in Fractured Rocks. EPA/
570/9-91-009. Office of Water, Washington, Gillham, R.W., E.A. Sudicky, J.A. Cherry, and
DC. 144 pp. E.O. Frind. 1984. An Advective-Diffusion
Concept to Solute Transport in Heterogeneous
Cohen, R.M. and J.W. Mercer. 1993. DNAPL Site Unconsolidated Geologic Deposits. Water
Evaluation. EPA/600/R-93/002 (NTIS PB93- Resour. Res. 20(3):369-378.
150217). R.S. Kerr Environmental Research
Laboratory, Ada, OK. [Also published by Gorelick, S.M., R.A. Freeze, D. Donohue, and
Lewis Publishers as C.K. Smoley edition, J.F. Keely. 1993. Groundwater Contamina-
Boca Raton, FL. 384 pp.] tion: Optimal Capture and Containment.
Lewis Publishers: Boca Raton, FL. 416 pp.

Haitjema, H.M., J. Wittman, V. Kelson, and N. Marquis, Jr., S. 1995. Dont Give Up on Pump
Bauch. 1994. WhAEM: Program Documenta- and Treat: Enhance It with Bioremediation.
tion for the Wellhead Analytic Element Soils & Groundwater Cleanup, August-
Model. EPA/600/R-94/210, 120 pp. Available September, pp. 46-50.
from EPA Center for Subsurface Modeling
Support, Ada, OK. [Includes Geographic McMurty, D.C., and R.O. Elton. 1985. New
Analytic Element Preprocessor (GAEP) and Approach to In-Situ Treatment of Contami-
Capture Zone Analytic Element Model nated Groundwaters. Environ. Progress
(CZAEM)] 4(3):168-170.

Haley, J.L., B. Hanson, C. Enfield, and J. Glass. National Research Council (NRC). 1994. Alterna-
1991. Evaluating the Effectiveness of Ground tives for Ground Water Cleanup. National
Water Extraction Systems. Ground Water Academy Press. 336 pp.
Monitoring Rev. 11(1):119-124. [Summary of Norris, R.D. et al. 1993. In-Situ Bioremediation
U.S. EPA (1989)] of Ground Water and Geological Material: A
Harvey, C.F., R. Haggerty, and S.M. Gorelick. Review of Technologies. EPA/600/R-93/124
1994. Aquifer Remediation: A Method for (NTIS PB93-215564). R.S. Kerr Environmen-
Estimating Mass Transfer Rate Coefficients tal Research Laboratory, Ada, OK. [13
and an Evaluation of Pulsed Pumping. Water authors; see also Norris et al., 1994]
Resour. Res. 30(7):1979-1991. Palmer, C.D. and W. Fish. 1992. Chemical
Hoffman, F. 1993. Ground-Water Remediation Enhancements to Pump-and-Treat Remedia-
Using Smart Pump and Treat. Ground tion. Ground Water Issue Paper. EPA/540/S-
Water 31(1):98-106. 92/001. R.S. Kerr Environmental Research
Laboratory, Ada, OK. 20 pp.
Keely, J.F. 1989. Performance Evaluation of
Pump-and-Treat Remediations. Superfund Papadopulos & Associates, Inc. and Conestoga-
Issue Paper. EPA/540/8-89/005. R.S. Kerr Rovers & Associates Ltd. 1993. Chem-Dyne
Environmental Research Laboratory, Ada, Site Trust Fund; 1992 Annual Report. Chem-
OK. 14 pp. Dyne Site, Hamilton, OH. April.

Knox, R.C., L.W. Canter, D.F. Kincannon, E.L. Piwoni, M.D. and J.W. Keeley. 1990. Basic
Stover, and C.H. Ward. 1984. State-of-the Art Concepts of Contaminant Sorption at Hazard-
of Aquifer Restoration. EPA/600/2-84/ ous Waste Sites. Ground Water Issue. EPA/
182a&b (National Technical Information 540/4-90/053. R.S. Kerr Environmental
Service [NTIS] PB85-181071 and PB85- Research Laboratory, Ada, OK. 7 pp.
181089). R.S. Kerr Environmental Research Rogers, L.L., R.U. Dowla, and V.M. Johnson.
Laboratory, Ada, OK. 1995. Optimal Field-Scale Groundwater
Mackay, D.M. and J.A. Cherry. 1989. Groundwa- Remediation Using Neural Networks and
ter Contamination: Pump-and-Treat Remedia- Genetic Algorithm. Environ. Sci. Technol.
tion. Environ. Sci. Technol. 23(6):630-636. 29(5):1145-1155.

Roy, W.R., I.G. Krapac, S.F.J. Chou, and R.A. Sites. EPA/540/G-88/003. Office of Solid
Griffin. 1992. Batch-Type Procedures for Waste and Emergency Response (OSWER)
Estimating Soil Adsorption of Chemicals. Directive 9283.1-2 (NTIS PB89-184618).
EPA/530/SW-87/006F (NTIS PB92-146190). Office of Solid Waste and Emergency Re-
Risk Reduction Engineering Laboratory, sponse, Washington, DC.
Cincinnati, OH. 100 pp.
U.S. Environmental Protection Agency (EPA).
Satkin, R.L. and P.B. Bedient. 1988. Effectiveness 1989. Evaluation of Ground-Water Extraction
of Various Aquifer Restoration Schemes Remedies: Volume 1, Summary Report (EPA/
Under Variable Hydrogeologic Conditions. 540/2-89/054, NTIS PB90-183583, 66 pp.);
Ground Water 26(4):488-498. Volume 2, Case Studies 1-19 (EPA/540/2-89/
054b); and Volume 3, General Site Data Base
Sims, J.L., J.M. Suflita, and H.H. Russell. 1992. Reports (EPA/540/2-89/054c). Office of Solid
In-Situ Bioremediation of Contaminated Waste and Emergency Response, Washington,
Ground Water. Ground Water Issue Paper. DC.
EPA/540/S-92/003. R.S. Kerr Environmental
Research Laboratory, Ada, OK. 11 pp. U.S. Environmental Protection Agency (EPA).
1991. Handbook: Stabilization Technologies
Starr, R.C. and J.A. Cherry. 1994. In Situ Reme- for RCRA Corrective Actions. EPA/625/6-91/
diation of Contaminated Ground Water: The 026. Center for Environmental Research
Funnel-and-Gate System. Ground Water Information, Cincinnati, OH. 62 pp.
U.S. Environmental Protection Agency (EPA).
Strack, O.D.L. et al. 1994. CZAEM Users Guide: 1992. Methods for Evaluating the Attainment
Modeling Capture Zones of Ground-Water of Cleanup Standards, Volume 2: Ground
Wells Using Analytic Elements. EPA/600/R- Water. EPA/230/R-92/014. Office of Solid
94/174, 58 pp. Available from EPA Center for Waste and Emergency Response, Washington,
Subsurface Modeling Support, Ada, OK. [See DC.
also, Haitjema et al., 1994]
U.S. Environmental Protection Agency (EPA).
Thomas, J.D., and C.H. Ward. 1989. In Situ 1993. Guidance for Evaluation the Technical
Biorestoration of Organic Contaminants in Impracticability of Ground-Water Restora-
the Subsurface. Environ. Sci. Technol. tion. EPA/540/R-93/080, OSWER 0234.2-25
23:760-786. (NTIS PB93-963507). Office of Solid Waste
Travis, C.C. and C.B. Doty. 1990. Can Contami- and Emergency Response, Washington, DC.
nated Aquifers at Superfund Site Be U.S. Environmental Protection Agency (EPA).
Remediated? Environ. Sci. Technol. 1994. Manual: Alternative Methods for Fluid
24(1):1464-1466. Deliver and Recovery. EPA/625/R-94/003.
U.S. Environmental Protection Agency (EPA). Center for Environmental Research Informa-
1988. Guidance on Remedial Actions for tion, Cincinnati, OH. 87 pp.
Contaminated Ground Water at Superfund

U.S. Environmental Protection Agency (EPA). Wilson, J.L., S.H. Conrad, W.R. Mason, W.
1995. Manual: Ground-Water and Leachate Peplinski and E. Hagen. 1990. Laboratory
Treatment Systems. EPA/625/R-94/005. Investigation of Residual Liquid Organics.
Center for Environmental Research Informa- EPA/600/6-90/004. R.S. Kerr Environmental
tion, Cincinnati, OH. 119 pp. Research Laboratory, Ada, OK. 267 pp.

9. EPA Publications Providing Further Information
The EPA publications listed below provide more McLean, J.E. and B.E. Bledsoe. 1992. Behavior
detailed information on the subjects discussed of Metals in Soils. Ground Water Issue. EPA/
in this document. Publications and additional 540/S-92/018. R.S. Kerr Environmental
copies of this brochure can be obtained at no Research Laboratory, Ada, OK. 25 pp.
charge (while supplies are available) from the
following sources: Palmer, C.D. and R.W. Puls. 1994. Natural
Attenuation of Hexavalent Chromium in
EPA/625-series documents: Office of Research Ground Water and Soils. Ground Water Issue.
and Development (ORD) Publications, P.O. EPA/540/S-94/505. R.S. Kerr Environmental
Box 19968, Cincinnati, OH 45219-0968; Research Laboratory, Ada, OK. 13 pp.
phone 513 569-7562, fax 513 569-7562.
Piwoni, M.D. and J.W. Keeley. 1990. Basic
Other EPA documents: National Center for Concepts of Contaminant Sorption at Hazard-
Environmental Publications and Information ous Waste Sites. Ground Water Issue. EPA/
(NCEPI), 11029 Kenwood Road, Cincinnati, 540/4-90/053. R.S. Kerr Environmental
OH 45242; fax 513 891-6685. Research Laboratory, Ada, OK. 7 pp.
Other documents, for which an NTIS acquisition Sims, J.L., J.M. Suflita, and H.H. Russell. 1991.
number is shown can be obtained from the Reductive Dehalogenation of Organic Con-
National Technical Information Service taminates in Soils and Ground Water. Ground
(NTIS), Springfield, VA 22161; 800 336- Water Issue. EPA/540/4-91/054. R.S. Kerr
4700, fax 703/321-8547. Environmental Research Laboratory, Ada,
OK. 12 pp.
Contaminant Transport and Fate
Huling, S.G. 1989. Facilitated Transport. Ground Wilson, J.L., S.H. Conrad, W.R. Mason, W.
Water Issue. EPA/540/4-89/003. R.S. Kerr Peplinski, and E. Hagen. 1990. Laboratory
Environmental Research Laboratory, Ada, Investigation of Residual Liquid Organics.
OK. 5 pp. EPA/600/6-90/004. R.S. Kerr Environmental
Research Laboratory, Ada, OK. 267 pp.
Huling, S.C. and J.W. Weaver. 1991. Dense
Nonaqueous Phase Liquids. Ground Water Site Characterization
Issue. EPA/540/4-91/002. R.S. Kerr Environ- U.S. Environmental Protection Agency (EPA).
mental Research Laboratory, Ada, OK. 21 pp. 1991. Site Characterization for Subsurface
Remediation. EPA/625/4-91/026. Center for

Environmental Research Information, Cincin- Environmental Research Laboratory, Ada,
nati, OH. 259 pp. OK. 14 pp.
U.S. Environmental Protection Agency (EPA). Mercer, J.W., D.C. Skipp, and D. Giffin. 1990.
1992. Estimating the Potential for the Occur- Basics of Pump-and-Treat Ground-Water
rence of DNAPL at Superfund Sites. OSWER Remediation Technology. EPA/600/8-90/003.
Publication 9355.4-07/FS. Office of Solid R.S. Kerr Environmental Research Labora-
Waste and Emergency Response, Washington, tory, Ada, OK. 58 pp.
Palmer, C.D. and W. Fish. 1992. Chemical
U.S. Environmental Protection Agency (EPA). Enhancements to Pump-and-Treat Remedia-
1993. Evaluation of the Likelihood of tion. Ground Water Issue Paper. EPA/540/S-
DNAPL Presence at NPL Sites. EPA/540/R- 92/001. R.S. Kerr Environmental Research
93/002 (OSWER 0355.4-13). Office of Solid Laboratory, Ada, OK. 20 pp.
Waste and Emergency Response, Washington,
DC. Repa, E. and D.P. Doerr. 1985. Leachate Plume
Management. EPA/540/2-85/004 (NTIS
U.S. Environmental Protection Agency (EPA). PB86-122330). Hazardous Waste Engineering
1993. Use of Airborne, Surface and Borehole Research Laboratory, Cincinnati, OH.
Geophysical Techniques at Contaminated
Sites: A Reference Guide. EPA/625/R-92/007. U.S. Environmental Protection Agency (EPA).
Center for Environmental Research Informa- 1989. Evaluation of Ground-Water Extraction
tion, Cincinnati, OH. Remedies: Volume 1, Summary Report (EPA/
540/2-89/054, NTIS PB90-183583, 66 pp.);
U.S. Environmental Protection Agency (EPA). Volume 2, Case Studies 1-19 (EPA/540/2-89/
1993. Subsurface Characterization and 054b); and Volume 3, General Site Data Base
Monitoring Techniques: A Desk Reference Reports (EPA/540/2-89/054c). Office of Solid
Guide; Vol. I: Solids and Ground Water; Vol. Waste and Emergency Response, Washington,
II: The Vadose Zone, Field Screening and DC.
Analytical Methods. EPA/625/R-93/003a&b.
Center for Environmental Research Informa- U.S. Environmental Protection Agency (EPA).
tion, Cincinnati, OH. 1992. Evaluation of Ground-Water Extraction
Remedies, Phase II. Oswer Publication
Pump-and-Treat Systems 9355.4-05, Vols. 1-2. Office of Solid Waste
Cohen, R.M., A.H. Vincent, J.W. Mercer, C.R. and Emergency Response, Washington, DC.
Faust, and C.P. Spalding. 1994. Methods for
U.S. Environmental Protection Agency (EPA).
Monitoring Pump-and-Treat Performance. 1992. General Methods for Remedial Opera-
EPA/600/R-94/123. R.S. Kerr Environmental tions Performance Evaluation. EPA/600/R-92/
Research Laboratory, Ada, OK. 102 pp. 002. R.S. Kerr Environmental Research
Keely, J.F. 1989. Performance Evaluation of Laboratory, Ada, OK. 37 pp.
Pump-and-Treat Remediations. Superfund U.S. Environmental Protection Agency (EPA).
Issue Paper. EPA 540/8-89/005. R.S. Kerr
1993. Guidance for Evaluating the Technical

Impracticability of Ground-Water Restora- McArdle, J.L., M.M. Arozarena, and W.E.
tion. EPA/540/R-93/080, OSWER 0234.2-25 Gallagher. 1987. A Handbook on Treatment
(NTIS PB93-963507). Office of Solid Waste of Hazardous Waste Leachate. EPA/600/8-87/
and Emergency Response, Washington, DC. 006 (NTIS PB87-152328). Hazardous Waste
Engineering Research Laboratory, Cincinnati,
U.S. Environmental Protection Agency (EPA). OH.
1993. Bioremediation Resource Guide. EPA/
542/B-93/004. Office of Solid Waste and U.S. Department of Defense Environmental
Emergency Response, Washington, DC. Technology Transfer Committee (DOD/
[Includes annotated list of more than 80 ETTC). 1994. Remediation Technologies
significant references] Screening Matrix and Reference Guide. EPA/
542/B-94/013 (NTIS PB95-104782). Office
U.S. Environmental Protection Agency (EPA). of Solid Waste and Emergency Response,
1994. Bioremediation in the Field. EPA/540/ Washington, DC.
N-94/501. Office of Solid Waste and Emer-
gency Response, Washington, DC. [Periodi- U.S. Environmental Protection Agency (EPA).
cally updated; latest issue No. 11, July, 1994] 1994. Ground-Water Treatment Technology
Resource Guide. EPA/542/B-94/009. Office
U.S. Environmental Protection Agency (EPA). of Solid Waste and Emergency Response,
1994. Manual: Alternative Methods for Fluid Washington, DC. [Includes annotated list of
Delivery and Recovery. EPA/625/R-94/003. more than 60 significant references]
Center for Environmental Research Informa-
tion, Cincinnati, OH. 87 pp. U.S. Environmental Protection Agency (EPA).
1994. Innovative Treatment Technologies
U.S. Environmental Protection Agency (EPA). Annual Status Report, 6th ed. EPA/542/R-94/
1995. In Situ Remediation Technology Status 005. Office of Solid Waste and Emergency
Report: Hydraulic and Pneumatic Fracturing. Response, Washington, DC.
EPA/542/K-94/005. Office of Solid Waste and
Emergency Response, Washington, DC. 15 U.S. Environmental Protection Agency (EPA).
pp. 1994. Superfund Innovative Technology
Evaluation Program: Technology Profiles, 7th
Ground-Water Treatment Methods ed. EPA/540/R-94/526. Risk Reduction
Canter, L.W. and R.C. Knox. 1986. Ground Water Engineering Laboratory, Cincinnati, OH. 499
Pollution Control. Lewis Publishers: Chelsea, pp.
MI. 526 pp. [Contains mostly same material
as Knox et al. (1984)] U.S. Environmental Protection Agency (EPA).
1995. Manual: Ground-Water and Leachate
Knox, R.C., L.W. Canter, D.F. Kincannon, E.L. Treatment Systems. EPA/625/R-94/005.
Stover, and C.H. Ward. 1984. State-of-the Art Center for Environmental Research Informa-
of Aquifer Restoration. EPA 600/2-84/ tion, Cincinnati, OH.
182a&b (NTIS PB85-181071 and PB85-
181089). R.S. Kerr Environmental Research
Laboratory, Ada, OK. [See also Canter and
Knox (1985)]

In Situ Ground-Water Treatment U.S. Environmental Protection Agency (EPA).
Norris, R.D. et al. 1993. In-Situ Bioremediation 1995e. In Situ Remediation Technology
of Ground Water and Geological Material: A Status Report: Treatment Walls. EPA/542/K-
Review of Technologies. EPA/600/R-93/124 94/004. Office of Solid Waste and Emergency
(NTIS PB93-215564). R.S. Kerr Environmen- Response, Washington, DC. 26 pp.
tal Research Laboratory, Ada, OK.[13 au-
thors; see also Norris et al., 1994]
Ground-Water Modeling
Bear, J., M.S. Beljin, and R.R. Ross. 1992.
Norris, R.D. et al. 1994. Handbook of Bioreme- Fundamentals of Ground-Water Modeling.
diation. Boca Raton, FL: Lewis Publishers. Ground Water Issue. EPA/540/S-92/005. R.S.
272 pp. [Contains same material as Norris et Kerr Environmental Research Laboratory,
al., 1993] Ada, OK. 11 pp.
Sims, J.L., J.M. Suflita, and H.H. Russell. 1992. Schmelling, S.G. and R.R. Ross. 1989. Contami-
In Situ Bioremediation of Contaminated nant Transport in Fractured Media: Models
Ground Water. Ground Water Issue Paper. for Decisionmakers. Ground Water Issue.
EPA/540/S-92/003. R.S. Kerr Environmental EPA/540/4-89/004. R.S. Kerr Environmental
Research Laboratory, Ada, OK. 11 pp. Research Laboratory, Ada, OK. 8 pp.
U.S. Environmental Protection Agency (EPA). U.S. Environmental Protection Agency (EPA).
1995a. In Situ Remediation Technology 1988. Selection Criteria for Mathematical
Status Report: Thermal Enhancements. EPA/ Models Used in Exposure Assessments:
542/K-94/009. Office of Solid Waste and Ground-Water Models. EPA/600/8-88/075
Emergency Response, Washington, DC. 22 (NTIS PB88-248752). Office of Health and
pp. Environmental Assessment, Washington, DC.
[Contains summary tables and descriptions of
U.S. Environmental Protection Agency (EPA). 63 analytical solutions and 49 analytical and
1995b. In Situ Remediation Technology numerical codes for evaluating ground-water
Status Report: Surfactant Enhancements. contaminant transport]
EPA/542/K-94/003. Office of Solid Waste and
Emergency Response, Washington, DC. 22 U.S. Environmental Protection Agency (EPA).
pp. 1994. Assessment Framework for Ground-
Water Model Applications. EPA/500/B-94/
U.S. Environmental Protection Agency (EPA). 003 (OSWER Directive 9029.00). Office of
1995c. In Situ Remediation Technology Solid Waste and Emergency Response,
Status Report: Cosolvents. EPA/542/K-94/ Washington, DC. 41 pp.
006. Office of Solid Waste and Emergency
Response, Washington, DC. 6 pp. van der Heijde, P.K.M. 1994. Identification and
Compilation of Unsaturated/Vadose Zone
U.S. Environmental Protection Agency (EPA). Models. EPA/600/R-94/028 (NTIS PB94-
1995d. In Situ Remediation Technology 157773). R.S. Kerr Environmental Research
Status Report: Electrokinetics. EPA/542/K- Laboratory, Ada, OK.
94/007. Office of Solid Waste and Emergency
Response, Washington, DC. 20 pp.

van der Heijde, P.K.M. and O.A. Einawawy. 1993. Ada, OK. [Summary information on models
Compilation of Ground-Water Models. EPA/ for porous media flow and transport,
600/R-93/118 (NTIS PB93-209401). R.S. hydrogeochemical models, stochastic models,
Kerr Environmental Research Laboratory, and fractured rock]