United States
Environmental Protection
Agency
Office of
Research and
Development
Office of Solid Waste
and Emergency
Response
EPA/540/S-95/500
July 1995
 &EPA     Ground  Water  Issue
                        Light Nonaqueous Phase Liquids
                        Charles J. Newell,* Steven D. Acree,** Randall R. Ross/
                        and Scott G. Huling**
The Regional Superfund Ground-Water Forum is a group of
scientists representing EPA's Regional Superfund Offices,
committed to the identification and resolution of ground-water
issues affecting the remediation of Superfund sites. Light
nonaqueous phase liquids (LNAPLs) have been identified by
the Forum as an issue of concern to decision makers. This
issue paper focuses on transport, fate, characterization, and
remediation of LNAPLs in the environment.

For further information contact Steve Acree (405) 436-8609,
Randall Ross (405) 436-8611, or Scott Huling (405) 436-8610
at RSKERL-Ada.

INTRODUCTION

Nonaqueous phase liquids (NAPLs) are hydrocarbons that
exist as a separate, immiscible phase when in contact with
water and/or air. Differences in the physical and chemical
properties of water and NAPL result in the formation of a
physical interface between the liquids which prevents the two
fluids from mixing. Nonaqueous phase liquids are typically
classified as either light nonaqueous phase liquids (LNAPLs)
which have densities less than that of water, or dense
nonaqueous phase liquids  (DNAPLs) which have densities
greater than that of water.  A previous issue paper developed
by the Robert S. Kerr Environmental Research Laboratory
reviews processes and management  issues pertaining to
DNAPLs (Huling and Weaver, 1991).

Light nonaqueous phase liquids affect ground-water quality at
many sites across the country. The most common LNAPL-
related ground-water contamination problems result from the
release of petroleum products. These products are typically
multicomponent organic mixtures composed of chemicals with
varying degrees of water solubility. Some additives (e.g.,
methyl tertiary-butyl ether and alcohols) are highly soluble.
Other components (e.g., benzene, toluene, ethylbenzene, and
xylenes) are slightly soluble. Many components (e.g., n-
dodecane and n-heptane) have relatively low water solubility
under ideal  conditions. Physical and chemical properties
which affect transport and fate of selected LNAPL compounds
and refined  petroleum products are presented in Table 1. In
general, LNAPLs represent potential long-term sources for
continued ground-water contamination at many sites.

LNAPL TRANSPORT THROUGH POROUS MEDIA

General Conceptual Model

Movement of LNAPLs in the subsurface is controlled by
several processes described in the following simplified
scenario (Figure 1).  Upon release to the environment, NAPL
(i.e., LNAPL or DNAPL) will  migrate downward under the
force of gravity. If a small volume of NAPL is released to the
subsurface, it will move through the unsaturated zone where a
fraction of the hydrocarbon will be retained by capillary forces
as residual  globules  in the soil pores, thereby depleting the
contiguous  NAPL mass until movement ceases. If sufficient
LNAPL is released, it will migrate until it encounters a physical
barrier (e.g., low permeability strata) or is affected by
buoyancy forces near the water table.  Once the capillary
fringe is reached, the LNAPL may move laterally as a
continuous, free-phase layer along the upper boundary of the
water-saturated zone due to gravity and capillary forces.
   Groundwater Services, Inc.
   Roberts. Kerr Environmental Research Laboratory.
         ^ I0'v •
                        Superfund Technology Support Center for
                        Ground Water

                        Robert S. Kerr Environmental
                        Research Laboratory
                        Ada, Oklahoma
                     Oflte* of soRrf Waste and i wr
                                           tft, SO
                     Wafter W, Kovaficfc* Jr,
                                                                                   Printed on Recycled Paper

-------
Table 1. Representative properties of selected LNAPL chemicals commonly found at Superfund sites (U.S.EPA, 1990) water and
        selected petroleum products (Lyman and Noonan, 1990)
to****
Methyl Ethyl Ketone
4-Methyl-2-Pentanone
Tetrahydrofuran
ftmsfyt
0.805
0.8017
0.8892
ttyftsunicf
Vi$tX>$lty
0.40
0.5848
0.55
Waterf
2.68 E+05
1 .9 E+04
3 E+05<1>
V^porf
fmmH0}
71.2
16
45.6 (2)
Hemy^ tawf
O»>$fc&tf
2.74 E-05 <2>
1 .55 E-04 (2)
1.1 E-04 <2>
  Benzene                    0.8765         0.6468
  Ethyl Benzene              0.867           0.678
  Styrene                     0.9060         0.751
  Toluene                    0.8669         0.58
  m-Xylene                   0.8642 <1>       0.608
  o-Xylene                    0.880 <1>         0.802
  p-Xylene                    0.8610 <1>       0.635

  Water                      0.998 <6>         1.14<6>

Common Petroleum Products
  Automotive gasoline         0.72-0.76 <3>     0.36-0.49 <3>
  #2 Fuel Oil                  0.87-0.95       1.15-1.97 <5>
  #6 Fuel Oil                  0.87-0.95       14.5-493.5  <">
  Jet Fuel (JP-4)              -0.75           -0.83 <5>
  Mineral Base
    Crankcase Oil             0.84-0.96 <6>     ~275<4>
                                                                 1.78
                                                                 1.52
                                                                 3
                                                                 5.15
                                                                 2
                                                                 1 .7
                                                                 1.98
                                                                       E+03
                                                                       E+02
                                                                       E+02
                                                                       E+02
                                                                       E+02
                                                                       E+02
                                                                       E+02<1>
76
7
5
22
9
7
9
   Values are given at 20°C unless noted.
   Value is at 25°C.
   Value is at unknown temperature but is assumed to be 20°- 30°C.
   Value is at 15.6°C.
   Value is at 38°C.
   Value is at 21 °C.
   Value is at 15°C.
5.43
7.9
2.28
6.61
6.91
4.94
7.01
E-03 <1>
E-03 <1>
E-03
E-03<1>
E-03 <1>
E-03 <1>
E-03 <1>
Although principal migration may be in the direction of the
maximum decrease in water-table elevation, some migration
may occur initially in other directions.  A large continuous-
phase LNAPL mass may hydrostatically depress the capillary
fringe and water table. Once the source is removed,
mounded  LNAPL migrates laterally, LNAPL hydrostatic
pressure is removed, and the water table eventually rebounds.
Infiltrating precipitation and passing ground water in contact
with residual or mobile LNAPL will dissolve soluble
components and form an aqueous-phase contaminant plume.
In addition, volatilization  may result  in further spreading of
contamination.

Contaminant Phase Distribution

LNAPL constituents may exist in any of four phases within the
subsurface.  The NAPL, aqueous, and gaseous phases were
mentioned above.  Contaminants may also partition to the
solid-phase material (i.e.,soil or aquifer materials).  In  the
unsaturated zone contaminants may exist in all four phases
                                                                           Dissolved
                                                                         Contaminants
                   Water Table
            LNAPL Contamination

           Ground-Water
         "*     Flow
        After Mercer and Cohen (1990)
                                                           Figure 1. Simplified conceptual model for LNAPL release and
                                                                    migration.

-------
                        Solid
        Soil Gas
                                          Water
                                            LNAPL
                                  After Huling & Weaver (1991)
Figure 2. Contamination in the unsaturated zone may be present
        in four physical states: gas, sorbed to soil materials,
        dissolved in water, or immiscible liquid.
                                                                      Water
                                  After DiGiulio and Cho (1990)
Figure 3. Partitioning of LNAPL among the four phases
        potentially found in the unsaturated zone.
(Figure 2).  In the saturated zone NAPL-related contaminants
may be present in the aqueous, solid, and NAPL phases.
NAPL constituents may partition, or move from one phase to
another, depending on environmental conditions (Figure 3).
For example, soluble components may dissolve from the
NAPL into passing ground water.  The same molecule may
adsorb onto a solid surface, and subsequently desorb into
passing ground water. The tendency for a contaminant to
partition from one  phase to another may be  described by
partition coefficients such as Henry's Law constant for
partitioning between water and soil gas. These  empirical
coefficients are dependent on the properties of the subsurface
materials and the NAPL. A clear understanding of the phase
distribution of contaminants is critical to evaluating remedial
decisions (Huling and Weaver, 1991). It is important to note
that this distribution is not static and may  vary over time due to
remedial actions and natural processes.

LNAPL Transport Parameters

Characteristics of the LNAPL and subsurface materials govern
transport at both the pore scale and field scale.  At the pore
scale, the following transport and fate parameters control
LNAPL migration and distribution.  At the  field scale, LNAPL
migration is much  more difficult to predict due to such factors
as complex release history and, most importantly, subsurface
heterogeneity.  However, the following  discussion of pore-
scale principles is  necessary for development of conceptual
models incorporating observations made  at  the field scale.  A
more detailed explanation of these concepts ( Mercer and
Cohen, 1990) and methods for measuring these properties
(Cohen and Mercer, 1993) are available in the literature.

Density

Density is defined as the mass of a substance per unit
volume. One way to express density of a fluid is the specific
gravity (S.G.) which is the ratio of the mass  of a given volume
of substance at a specified temperature to the mass of the
same volume of water at the same temperature. If a NAPL
has an S.G. less than water, generally  less than 1.0, it is less
dense than water (i.e., LNAPL) and will float on  water.  If it has
an S.G. greater than water, generally greater than 1.0, it is
denser than water (DNAPL). The density of most fluids
generally decreases as temperature increases.
Consequently, the density of fluids considered to be DNAPLs
under normal subsurface conditions may decrease during
remedial actions which impart heat to the subsurface (e.g.,
Johnson and Leuschner, 1992). A decrease in density of
DNAPLs which have densities near that of water (e.g., some
coal tar residues) may result in sufficient reduction to
temporarily convert the DNAPL to an LNAPL.  Density not only
affects the buoyancy of a liquid but also the  subsurface
mobility. The hydraulic conductivity of a porous medium is a
function of the density and viscosity of the liquid. As the
density increases, the hydraulic conductivity with respect to
the liquid also increases.

Viscosity

Viscosity is the resistance of a fluid to flow.  Dynamic, or
absolute, viscosity is expressed in units of mass per unit
length per unit time. This resistance is also temperature
dependent.  The viscosity of most fluids will decrease as the
temperature increases.  The lower the viscosity, the less
energy required for a fluid to flow in a porous medium. The
hydraulic conductivity increases as the fluid viscosity
decreases.

Interfacial Tension

When two liquids, which are immiscible, are in contact, an
interfacial energy exists between the fluids, resulting in a
physical interface. Interfacial tension is the surface energy at
the interface that results from differences in the forces of
molecular attraction within the fluids and at the interface
(Bear, 1972).  It is expressed in units of energy per unit area.
In general, the greater the interfacial tension, the greater the
stability of the interface between the liquids. Interfacial
tension is affected by temperature (Davis and  Lien, 1993),
changes in pH and the presence of surfactants and dissolved
gases.  It is an important factor affecting wettability (Mercer
and Cohen,  1990).

Wettability

Wettability is generally defined as the overall tendency of one
fluid to spread on or adhere to a solid surface  (i.e.,

-------
preferentially coat) in the presence of another fluid with which
it is immiscible. This concept has been used to describe fluid
distribution at the pore scale.  In a multiphase system, the
wetting fluid will preferentially coat (wet) the solid surfaces and
tend to occupy smaller pore spaces. The non-wetting fluid will
generally be restricted to the largest interconnected pore
spaces. In the vadose zone, where air, water, and LNAPL are
present, liquids, usually water, preferentially wet solid
surfaces.  However, under conditions where only LNAPL and
air are present, LNAPL will preferentially coat the mineral
surfaces and displace air from pore spaces. In the saturated
zone, with only water and LNAPL present, water will generally
be the wetting fluid and will displace LNAPL from pore spaces.
Wettability is affected by such factors as NAPL and aqueous-
phase composition, presence of organic matter, surfactants,
mineralogy, and saturation history of the porous medium
(Mercer and Cohen, 1990). Some researchers have
concluded that wetting of subsurface media by NAPL may be
heterogeneous due to subsurface variability and the many
factors that influence wettability (Anderson, 1986). In
summary, wettability is a qualitative indicator useful to
understand the general behavior of NAPLs in multiphase
systems and has been used extensively in the petroleum
industry (Anderson, 1986). Actual wettability measurements
of NAPLs on solid surfaces are usually reported for flat,
homogeneous material which is  unrepresentative of complex
aquifer and soil material.  Refined petroleum products typically
found at Superfund sites can generally be considered the non-
wetting fluid in water:LNAPL systems, and the wetting fluid in
LNAPL:air systems.

Capillary Pressure

Capillary pressure  is the  pressure difference across the
interface between the wetting and non-wetting phases and is
often expressed as the height of an equivalent water column.
It determines the size of the pores in which an interface can
exist.  It is a measure of the relative attraction of the molecules
of a liquid (cohesion) for each other and for a solid surface
(adhesion).  Capillary pressure is represented by the tendency
of the porous medium to attract the wetting fluid and repel the
non-wetting fluid  (Bear, 1972). The capillary pressure of the
largest pore spaces must be exceeded before the  non-wetting
fluid  (generally NAPL) can enter the porous medium.  The
minimum pressure required for the NAPL to enter the medium
is termed the entry pressure.

In  general, capillary pressure increases with decreasing pore
size, decreasing initial moisture content, and increasing
interracial tension.  Capillary conditions affect the
configuration and magnitude of trapped residual NAPL. Field
observations of the effects of capillary pressure include
preferential LNAPL migration through coarse-grained
materials (e.g., sands and gravels), rather than fine-grained
materials (e.g., silts and clays). Analytical expressions
describing relationships between capillary pressure and NAPL
movement under hydrostatic and hydrodynamic conditions are
compiled by Mercer and Cohen (1990). Although these
approximate expressions do not account for complicated  pore
geometries and distributions present in most systems,
evaluation of site conditions using such expressions may often
be useful in refining the conceptual model for NAPL transport.

Although the capillary forces that hold residual NAPL in pores
are relatively strong, they can be overcome to some degree
by viscous forces associated with ground-water flow.
However, complete mobilization of residual hydrocarbons is
very difficult or impossible to achieve in most aquifers by
manipulating hydraulic gradient alone (Wilson and Conrad,
1984). The required hydraulic gradients are so high for many
aquifers (greater than 1 ft/ft) that no reasonable configuration
of pumping and injection wells could sweep all of the residual
NAPL trapped in the pores of the aquifer.

Saturation and Residual Saturation

Saturation is the relative fraction of total pore space containing
a particular fluid (e.g., NAPL) in a representative volume of a
porous medium.  The mobility of an LNAPL is related to its
saturation in the medium as described by the relative
permeability function discussed below.  The saturation level
where a continuous NAPL becomes discontinuous and is
immobilized by capillary forces is known as the residual
saturation (Sr). Residual saturation of LNAPL represents a
potential source for continued ground-water contamination
that is tightly held in the pore spaces and not readily removed
using currently available remediation technologies. The
magnitude of residual saturation is affected by several factors
including pore-size distribution, wetting properties of the fluids
and soil solids, interfacial tension, hydraulic gradients, ratios
of fluid viscosities and densities, gravity, buoyancy forces, and
flow rates (Mercer and Cohen, 1990; Demond and Roberts,
1991). Due to the known heterogeneity of subsurface
systems with regard to these factors, it follows that residual
saturation in the subsurface is also highly variable.

Data compiled by Mercer and Cohen (1990) indicate the
residual saturation of most NAPLs in these studies ranged
from about 10% to 20% in the unsaturated zone  and about
15% to 50% of the total pore volume in the saturated zone.
The potential for higher retention of NAPLs in the saturated
zone than in the unsaturated zone is due to several factors
including: 1) potential existence of the NAPL as the wetting
fluid relative to air in the unsaturated zone resulting in NAPL
spreading to adjacent pores with  residual held in small pore
spaces, 2) existence of the  NAPL as the non-wetting fluid in
the saturated zone resulting in NAPL present as blobs in
larger pore spaces, and 3) the relatively high fluid density ratio
of NAPL to air in the vadose zone resulting in drainage
(Anderson, 1988).

Relative Permeability

Relative permeability is the  ratio of the effective permeability
of the medium to a fluid at a specified saturation  and the
permeability of the medium  to the fluid at 100% saturation.
Values for relative permeability range between 0 and 1. A
simplified relative  permeability diagram for a hypothetical
LNAPL/water system (Figure 4) illustrates how two fluids
interfere with each other to reduce mobility. Similar, yet more
complex relationships exist  in the unsaturated zone where
three fluids (air, water, and NAPL) may be present (van Dam,
1967; Ferrand et al., 1989).  At most points on the curves, the
relative permeabilities of NAPL and water do not sum to one
because interference reduces the overall mobility of both
fluids in the porous medium. The curves also illustrate that a
minimum saturation must be attained before the permeability
to a fluid is non-zero (Schwille, 1988). The minimum
saturation for the wetting fluid has been termed irreducible
saturation (Sjr) and for the non-wetting fluid, generally NAPL,

-------
has been termed residual saturation. It should be noted that
the example described above is highly simplified for purposes
of this discussion. In reality, an infinite set of curves, bounded
by main curves for drainage and imbibition, describe the
relative permeability function.

Relative permeability curves (Figure 4) can be used to
describe different types of multiphase flow regimes, all of
which may exist at any particular site (Williams and Wilder,
1971):

Zone I:  LNAPL occurs as a potentially mobile, continuous
        phase and saturation is high. Water is restricted to
        small pores. The relative permeability of water is
        bw. Such conditions may be observed within large
        mobile product accumulations.

Zone II: Both LNAPL and water occur as continuous phases,
        but, generally, do not share the same pore spaces.
        However, the relative permeability of each fluid is
        greatly reduced by the saturation of the other fluid.
        Such conditions may be representative of zones of
        smaller mobile product accumulations at the water
        table.

Zone III: LNAPL is discontinuous and trapped as residual in
        isolated pores.  Flow is almost exclusively the
        movement of water, not  LNAPL. Examples of such
        conditions may be found within zones of residual
        LNAPL retained below the water table.
       1.00
                 Kr = relative permeability
    o.
    <
       0.10-
       0.01
LNAPL
 Flow
                            Mixed Flow
                                                 0.10
                                                     £
                                                     k_
                                                    ^
                                                 0.01
         or-
                        Water Saturation
                                               H 100%
                       LNAPL Saturation
                                             -H o
                After Williams and Wilder (1971)
Figure 4.  Hypothetical relative permeability curves for water and
        an LNAPL in a porous medium.
                                              LNAPL Migration at the Field Scale

                                              Darcy's Law

                                              Various forms of Darcy's Law may be used to describe fluid
                                              migration in porous media under many conditions. For
                                              example, a form of Darcy's Law, including relative permeability
                                              as a function of saturation, may be used to describe LNAPL
                                              flow when LNAPL saturation is less than 100%. However,
                                              movement of LNAPL can be examined using a simple
                                              conceptual model to understand the effects of LNAPL physical
                                              properties on mobility. The one-dimensional migration of
                                              LNAPL in  an LNAPL-saturated system can be represented
                                              using the following form:
                                                      v  » -(k p  g/jL)  (dh/dl)
                                                 (1)
                                              where
                                                      v = Darcy velocity (L/T)
                                                      k = intrinsic permeability (L2)
                                                      p = density of NAPL  (M/L3)
                                                      g = force of gravity (L/T2)
                                                      p. = dynamic (absolute) viscosity  (M/L*T)
                                                      dh/dl = hydraulic gradient of NAPL mass (LA)
Hydraulic conductivity of the medium is proportional to the
density and inversely proportional to the viscosity of the
LNAPL. The potential mobility of various fluids may be
compared using the ratio of density to viscosity. High ratios
correspond to greater potential mobility. In the subsurface,
LNAPLS are subjected to biotic and abiotic (volatilization and
solubilization) weathering processes which change the
composition (mass fraction) of product (Johnson et al.,
1990b). The weathering process may alter the overall LNAPL
properties, such as increasing the viscosity. A significant
change in these properties will affect the potential mobility of
the LNAPL

Field Scale Versus Pore Scale

While LNAPL migration at the pore scale can be described
using some of the physical relationships presented above,
migration and distribution of LNAPL at the field scale is
controlled by a complex combination of release factors, soil/
aquifer properties, and LNAPL characteristics (Mercer and
Cohen, 1990) including:

     •  volume of LNAPL released;
     •  release rate (e.g., one-time "slug" event vs. long-term
        continual discharge);
     •  LNAPL infiltration area at the release site;
     •  properties of the LNAPL (e.g., density, viscosity);
     •  properties of the soil/aquifer media (e.g.,
        permeability, pore size distribution);
     •  fluid/porous media relationships (e.g., wettability);
     •  lithology and stratigraphy; and
     •  macro-scale features (e.g., fractures, root holes).

To illustrate how these factors can combine to control LNAPL
migration, a series of conceptual models is presented below
which describe a general LNAPL release scenario. These
conceptual models are intended to convey several

-------
 fundamental principles. However, actual release scenarios
 are strongly affected by numerous site-specific parameters
 that may not necessarily be represented in this discussion.

 LNAPL Migration Through Vadose Zone

 After release on the surface, LNAPL moves vertically
 downward under the force of gravity.  For small volume
 LNAPL releases, all of the LNAPL may eventually be retained
 in pores and fractures in the unsaturated zone (Figure 5).
 Infiltration of water through the residual LNAPL due to
 recharge or increased water-table elevations slowly dissolves
 soluble constituents, resulting in an aqueous-phase
 contaminant plume. Migration of vapors may also spread
 contamination (Mendozaand McAlary, 1989). LNAPL
 migration is influenced by heterogeneity in subsurface media
 and may be complex. For example, it may preferentially
 migrate laterally through more permeable pathways or
 accumulate and migrate along low permeability layers above
 the water table (Figure 6).

 Accumulation at the Water Table

 If a sufficient volume of  LNAPL is released, it may migrate
 through the unsaturated zone toward the water table and zone
 of water held by capillary forces above the water table (i.e.,
 capillary fringe). As LNAPL approaches the water table,
 entering regions of increasing water saturation, it may migrate
 laterally. Lateral migration is controlled by the LNAPL head
 distribution. In general, migration may be expected to be
 greatest in the direction of ground-water flow (i.e., maximum
 decrease in water-table elevation).  However, migration may
 occur initially in other directions in response to the hydraulic
 gradients induced by an LNAPL mound (Figure 1).  A
 relatively large LNAPL accumulation may result in
 compression or collapse of the capillary fringe and, potentially,
 depression of the water table.
                                   Residual Saturation of
                                      LNAPL in Soil
                                         Ground-Water
                                     •*	Flow

               After Waterloo Centre for Groundwater Research (1989)
Figure 5. LNAPL is retained at residual saturation in unsaturated
        zone.  Leaching by infiltrating water and migration of
        hydrocarbon vapors result in ground-water
        contamination.
                            LNAPL
                            Release
           Low
           Permeability
           Units
                After Waterloo Centre for Groundwater Research (1989)
Figure 6. LNAPL affected by low permeability units prior to
         reaching water table.


Laboratory column experiments (Abdul, 1988) indicate LNAPL
may migrate freely through the upper region of the vadose
zone where water is present at irreducible saturation.
Movement through regions of increasing water saturation at
the capillary fringe depends on displacement of water from
these zones. In general, increasing LNAPL head is required
to displace water with increasing depth within these zones.
Rate and volume of LNAPL release and capillary properties of
subsurface materials are factors that affect the depth of
LNAPL penetration within the  water-saturated regions (Abdul,
1988).  In general, increases in release rates and volumes
lead to increased LNAPL head and increased depth of
penetration. Increases in capillary forces associated with fine
grained materials result in  increased LNAPL head required to
penetrate and displace pore water.

The conceptual model presented above is a relatively
simplified representation of potential subsurface conditions.
As discussed by Farr et al. (1990)  and Lenhard and Parker
(1990), sharp interfaces between zones saturated with
LNAPL, water, and air generally do not exist in the subsurface
at most sites. Porous media above the water-saturated zone
are generally filled with varying saturations of LNAPL, water,
and/or air (Figure 7).  Thus, the conceptual model often
reported in the literature indicating  a discrete LNAPL-water
interface should not be assumed.  Instead, it is reasonable to
assume that the relative saturation ranges reported for the
LNAPL and water phases in the mixed flow region (i.e., region
II) of Figure 4 are more representative of actual LNAPL-water
interface areas. From a remediation perspective, it is
apparent that LNAPL mobility  is  compromised due to the
water saturation in this interface area.

LNAPL Smearing Due to  Fluctuating Water Table

Accumulations of LNAPL at or near the water table are
susceptible to "smearing" from changes in  water-table
elevation such as those that occur  due to seasonal changes in
recharge/discharge or tidal influence  in coastal environments.
Mobile LNAPL floating above the water-saturated zone will
move vertically as the ground-water elevation fluctuates
(Figure 8). As the water table  rises or falls, LNAPL will be

-------
               Water and Air
           UNAPU water an* Air
           k V»s
              Porous Medium
                                                LNAPL
                                                IVater
                                                Well
                              After Farretal. (1990)
Figure 7. Conceptualization of a multiphase fluid distribution in
        porous medium and monitoring well screened within
        the medium.
retained in the soil pores, leaving behind a residual LNAPL
"smear zone".  If smearing occurs during a decline in ground-
water elevations, residual LNAPL may be trapped below the
water table when ground-water elevations rise.

A similar situation may develop during product recovery
efforts. LNAPL will flow towards a recovery well or trench in
response to the gradient induced by water-table depression.
LNAPL residual will be retained below the water table as the
water-table elevation returns to pre-pumping conditions.

LNAPL Migration in Fractured Media

LNAPL introduced into an unsaturated, fractured medium
follows a complex pathway based on the characteristics of the
rock, such as fracture density, orientation, and aperture
distribution (Figure 9).  The degree of fracture interconnectivity
may not be definable due to the extreme heterogeneity of
most fractured systems and the lack of economical aquifer
characterization technologies (U.S.EPA, 1992a). Many clay
units,  once considered to be impermeable, often act as
fractured media with preferential pathways for vertical and
horizontal NAPL migration. Although this concern is
particularly true for DNAPL, it is applicable to LNAPL in
unsaturated systems.
                    Prior to Rise in Water Table
        Following Rise in Water Table
                    LNAPL trapped by capillary forces

                    Mobile LNAPL
                              After API (1989)
Figure 8. Effect of rising water table on LNAPL distribution in porous medium.  A similar effect may be seen with a falling water table.

-------
                After Waterloo Centre for Groundwater Research (1989)
 Figure 9. Potential LNAPL migration in fractured medium.
 LNAPL Migration Through Man-Made Pathways

 LNAPL will move through man-made preferential pathways,
 such as improperly grouted monitoring wells or trenches
 containing distribution piping or utilities.  Trenches backfilled
 with pea gravel or other coarse-grained material can provide a
 horizontal pathway for LNAPL migration in the near-surface
 environment.
 FATE OF LNAPLs IN THE SUBSURFACE

 As illustrated in Figure 3, organic compounds which compose
 NAPLs may partition among four separate phases in the
 subsurface: gaseous, solid, aqueous, and NAPL. The
 subsurface fate of chemicals in these phases is largely
 determined by volatilization, dissolution, sorption, and
 degradation  processes.

 Volatilization

 Volatilization of organic compounds occurs by two primary
 pathways: volatilization from water and NAPL. Henry's Law
 describes the partitioning of an organic compound between
 the aqueous and gaseous phases. For a dilute solution,
 concentrations less than approximately 103moles/l, the ideal
 gas vapor pressure of a volatile solute is proportional to its
 mole fraction in solution. More simply stated, the escaping
 tendency of the solute molecules from the water phase to the
 air phase is proportional to the concentration in the water.
 This relationship assumes local equilibrium between water
 and air and is useful for estimating the potential for organic
 chemical transport from water to  air, and from hydrocarbon
 vapors to water (Mendoza and McAlary, 1989).  For more
 concentrated solutions or pure phase compounds,
 volatilization  is best described by Raoult's Law, which states
 that the vapor pressure over a solution is equal to the mole
fraction of the solute times the vapor pressure of the pure
 phase liquid. Again, this is a measure of the escaping
tendency of molecules from the NAPL to the gaseous phase.

 Henry's Law and Raoult's Law are useful to describe the
partitioning of organic compounds between fluid and gaseous
 phases when the system is at equilibrium. Most volatilization
 calculations are dependent on the assumption that the LNAPL
 and air are in chemical equilibrium (Johnson et al., 1990a).
 However, it is reasonable to assume that equilibrium
 conditions do not occur in many volatilization-based
 subsurface remediation systems. The actual transfer process
 in a dynamic and complex subsurface system must take into
 account numerous parameters that are not included in these
 general partitioning equations. A practical approach to
 evaluating soil ventilating systems involves an in-depth
 analysis of numerous site-specific parameters and processes
 (Johnson et al., 1990b).

 Dissolution

 A NAPL in physical contact with ground water will dissolve
 (solubilize, partition) into the aqueous phase. The solubility of
 an organic compound is the equilibrium concentration of the
 compound in water at a specified temperature and pressure.
 For all practical purposes, the solubility represents the
 maximum concentration of that compound in water. The
 solubilities of the compounds most commonly found at
 Superfund sites range over several orders of magnitude.
 Several parameters affecting solubility include temperature,
 pH, cosolvents, dissolved organic matter, and dissolved
 inorganic compounds (salinity).

 For a multicomponent NAPL in contact with water, the
 equilibrium dissolved-phase concentrations may be estimated
 using the solubility of the pure liquid in water and its mole
 fraction in the NAPL mixture (Feenstra et al., 1991). The
 maximum concentration that can be achieved in this scenario
 is referred to  as the effective solubility, as indicated in
 Equation 2.
                     KS,
(2)
where
        S" = effective aqueous solubility of compound i in
             NAPL mixture
        X{ = mole fraction of compound i in NAPL mixture
        Si = aqueous solubility of the pure-phase compound

Two examples are presented which illustrate the effect of
dissolution on multicomponent LNAPLs. First, consider a two-
component NAPL with equal mole fraction (0.5), where the
solubilities of the pure-phase compounds are 1000 mg/l and
10 mg/l, respectively, and the effective solubilities are 500
mg/l and 5 mg/l, respectively. The  more soluble compound
will potentially partition into ground  water 100-fold more
readily than the less soluble compound.  According, less
soluble compounds will primarily be associated with the NAPL
phase  and dissolution and transport in the aqueous phase will
be limited  relative to more soluble components. Second,
consider a "gasoline" containing numerous compounds
(Johnson et al., 1990b) where the mole  fractions of benzene
were 0.0076 and 0.0021 for fresh and weathered gasoline,
respectively. The solubility for benzene is 1780 mg/l
(U.S.EPA, 1990). Yet, the predicted effective solubility would
be only 13.5 mg/l and 3.78 mg/l for fresh and weathered
gasoline, respectively.

-------
The effective solubility represents the concentration that may
occur at equilibrium under ideal conditions.  Laboratory
studies (Banerjee, 1984) indicate that effective solubilities
calculated using Equation 2 are reasonable approximations
for mixtures of organic liquids that are hydrophobia,
structurally similar, and have low solubilities. Effective
solubilities of components in more complex mixtures, such as
petroleum products, appear to be in error by no more than a
factor of two (Leinonen and Mackay, 1973). However, the
degree to which these compounds partition to the water phase
is a function of many variables including cosolvency (Rao et
al., 1991). Cosolvency effects may occur in cases where
dissolution of  highly soluble components (e.g., alcohols)
significantly increase the solubility of other components.

In general, higher dissolution rates may be associated with
higher ground-water velocities, higher LINAPL saturation in the
subsurface, increased contact area between LNAPL and
water, and LNAPLs with a high fraction of soluble components
(Mercer and Cohen, 1990; Miller et al., 1990). However,
recent studies (e.g., Powers et al., 1992) indicate that non-
equilibrium effects may limit contaminant mass transfer by
dissolution under certain conditions such as high ground-
water velocity. Laboratory and modeling studies conducted by
many researchers (e.g., Borden and Kao, 1992; Geller and
Hunt, 1993; Powers et al., 1991) indicate that complete
dissolution of  an LNAPL may require hundreds or thousands
of pore volumes of water under ideal field conditions. These
studies observed initially high aqueous contaminant
concentrations which  were followed by a period of rapid
decline and an asymptotic period  during which concentrations
declined slowly.

Sorption

Sorption is defined as the interaction of a contaminant with a
solid (Piwoni and Keeley, 1990).  In soil or aquifer material
contaminated with LNAPL, contaminants from the LNAPL will
partition onto  solid phase material. The primary pathway in
which this process occurs is through the water phase, as
indicated in Figure 3.  For example, when LNAPL is released
into the subsurface, components will dissolve into the
aqueous phase, then partition onto aquifer material.
Numerous parameters affect sorption at hazardous waste
sites including solubility, polarity, ionic charge, pH, redox
potential, and the octanol/water partition coefficient (Piwoni
and Keeley, 1990).

In general, solid-phase (adsorbed) contaminants may
represent a small fraction of the total contaminant mass in soil
and aquifer material where continuous phase or residual
NAPL exists.  The majority of the  contaminant mass in these
systems is typically present in the nonaqueous liquid phase.
Desorption of the contaminant (i.e. mass transfer of
contaminant from the solid phase to the water phase) is often
a rate-limited  step and is partially responsible for the tailing
effect commonly observed in ground-water pump-and-lreat
systems.

Analytical results from an LNAPL-contaminated  aquifer or  soil
sample extracted and analyzed in the laboratory represent the
total contaminant mass associated with the sample. Careful
evaluation of the chemical  and physical properties of the
contaminant(s) and soil/aquifer sample is necessary to
evaluate the phase(s) in which the contaminant exists. A
practical approach has been developed to evalgate
contaminant phase distribution by applying equilibrium
partitioning theory (Feenstra et al., 1991).

Biodegradation

Many of the LNAPL-related compounds are amenable to
biological degradation in the aqueous phase by naturally
occurring microorganisms in the subsurface. However, there
is an important distinction between aqueous-phase and  NAPL
biodegradation. The distinction is the inability to create and
maintain conditions that are conducive to microbial activity
within a NAPL.  In brief, biodegradation of pure phase
hydrocarbon does not appear to be practical and has not been
demonstrated. Considerable research has focused on
evaluating aerobic and anaerobic biodegradation and
transformation processes. These processes play an important
role in the ultimate fate of LNAPLs in the subsurface, both in
the form of naturally occurring and actively engineered
remediation processes (Norris et al.,  1994).

LNAPL SITE CHARACTERIZATION

Success or failure of an LNAPL remediation program depends
in large measure on the remediation objectives and adequacy
of the site characterization. This section will focus on issues
pertaining to LNAPL investigations. Discussion of strategies
and techniques for characterization of aqueous-phase and
sorbed contamination will be limited.  More information
regarding  site characterization concepts, techniques, and
strategies is  provided by U.S.EPA (1991) and Cohen and
Mercer (1993).

The focus of site characterization  is often to provide
information necessary to define and evaluate potential
remedial options. Specific objectives may include
determination/delineation of subsurface contamination in the
aqueous, gaseous, solid, and LNAPL phases; mobile and
residual LNAPL; migration rates/directions of the mobile
phases; geologic controls on LNAPL movement; LNAPL
properties; and other pertinent fluid/media properties. The
level of detail and the type of data required will be  site-
specific, partially dictated by the remedial technologies under
consideration and practical economic constraints, and often
restricted by available characterization technologies.  Such
limitations include lack of practicable methods for detailed
delineation of many parameters of interest including  LNAPL
distribution and saturation and hydraulic conductivity
distributions.

Conceptual Model

The conceptual model is the interpretation and assimilation of
all site-related information into assumptions and hypotheses
regarding  contaminant sources, subsurface contaminant
distribution, and dominant transport/fate processes.  The basis
for this conceptual model may include one or more of the
conceptual models discussed in the previous section.

Characterization is best conducted in a phased approach,
beginning with a review of site history, contaminant properties,
and regional/local studies.  This review should include the
following (API, 1989):

-------
       1.  Information on storage, transportation, use,
          monitoring, and disposal of LNAPLs at the site;
       2.  Locations, volumes, and timing of any known LNAPL
          releases;
       3.  Locations of underground piping, structures, or
          utilities which might influence LNAPL flow;
       4.  Regional/local geologic and hydrogeologic studies,
          soil  surveys, climatic data, and pertinent maps/
          historic photographs of the site; and
       5.  Preliminary information, available from the literature,
          concerning pertinent contaminant transport and fate
          parameters for the site-specific contaminants.

 An initial conceptual model for contaminant distribution,
 transport, and fate at the site is then formulated and used to
 plan further characterization. Each characterization phase  is
 designed to test and refine this model. The iterative process
 continues throughout remedial design and during remedial
 operations.

 Soil/Aquifer Material

 Lrthologic, stratigraphic, and structural features may often be
 dominant influences on LNAPL movement.  As expected,
 LNAPLs migrate faster through relatively permeable features
 (e.g.,  root holes, fractures, sandy layers, utility trenches, etc.)
 in the unsaturated zone than through less permeable
 materials. Characterization of subsurface heterogeneity to the
 extent practicable  may provide valuable information
 concerning potential LNAPL migration and distribution.
 Analyses of geologic materials exposed in trenches and
 obtained from borings may be used at some sites to improve
 the conceptual model and identify features that potentially
 serve  as preferential pathways for LNAPL.  Non-invasive
 geophysical methods may also provide information regarding
 the distribution of geologic materials and stratigraphic or
 structural features  that may control contaminant migration.
 However, detailed  characterization of heterogeneity and
 LNAPL distribution often will not be possible due to
 subsurface complexity and limitations in characterization tools
 and techniques (U.S.EPA, 1992a).

 Fractured rock and karstic settings may represent extreme
 examples of heterogeneous environments.  These settings
 pose exceptionally difficult problems in site characterization
 due to the complexity of transport pathways. Efficient
 techniques for characterizing contaminant transport (i.e.,
 LNAPL, aqueous phase, and gaseous phase) at such sites
 are not available currently.

 Hazards of Invasive Characterization Methods

 Certain hazards  may exist during invasive characterization at
 LNAPL contamination sites. The potential for increasing the
 vertical extent of contamination should be considered when
 evaluating drilling and well installation programs. Drilling
 through mobile LNAPL (e.g., liquids perched on low
 permeability units above the water table) may result in
 contaminating deeper intervals.  Such risks  may be reduced
 using appropriate techniques such as detailed observation
 and screening of geologic materials.  The risk of explosion or
fire (API, 1989) may exist at sites contaminated with
flammable materials (e.g., most liquid petroleum products).
  During drilling operations, LNAPL, contaminated soil and
  aquifer material, and vapors are brought to the surface where
  conditions for ignition may exist.  LNAPL is not only a concern
  as a fire and explosion hazard, but also with regard to
  chemical exposure to the driller and sampling crews.
  Monitoring and  mitigation of explosion and exposure risks
  should be considered when planning field operations.

  Information from Borings and Excavations

  Multiple borings with continuous sampling of subsurface
  media or excavation of pits/trenches will generally be required
  to define stratigraphy and estimate properties of the
  subsurface media and fluids. Media properties useful in
  contaminant fate and transport studies at NAPL sites include
  texture, porosity, permeability, organic carbon content,
  isotropy,  and  heterogeneity. Estimation of fluid/media
  properties, such as fluid saturation and capillary pressure-
  saturation relationships, may also be useful. Such data may
  provide the basis for evaluating potential LNAPL distribution
  and mobility at the macro scale. Various laboratory methods
 for measuring such properties are discussed in Cohen and
  Mercer (1993).  Applicability of these methods will be site
 specific.  Hydrogeologic information, including the depth to
 ground water, delineation of perched ground water or LNAPL,
 hydraulic gradient, and  hydraulic conductivity of saturated
 materials, is also essential.  This information may be used to
 evaluate ground-water flow directions  which will also be the
 potential flow directions for the LNAPL. Characterization of
 temporal as well as spatial variations in many of these
 parameters generally will be required. For example,
 information on seasonal variations in water-table elevations
 may aid in evaluating mobile LNAPL flow directions and
 LNAPL distribution above and below the average water-table
 elevation. Installation of wells will be required to obtain much
 of this information. These data will provide the basis for
 evaluating LNAPL mobility on the site  scale.

 Methods for direct detection  of NAPL in soil samples are
 relatively limited. Visual observation often has been relied
 upon to make this determination in the field.  However, it may
 be difficult or impossible to observe NAPLs which are
 colorless or clear, heterogeneously distributed in the sample,
 or present at low saturations (Huling and Weaver, 1991).

 Recently,  qualitative  techniques to enhance rapid, visual
 observations of NAPL in drill cuttings or cores have been
 evaluated (Cohen et  al.  1992).  Centrifugation, examination
 under ultraviolet  light, soil-water separation tests, and addition
 of hydrophobic dye to preferentially stain NAPL (i.e., Sudan
 IV) were studied. Of these methods, the use of hydrophobic
 dye and ultraviolet light examination for fluorescent NAPLs
 were found to be the simplest and most effective methods.
 These methods provide  only limited, qualitative information
 concerning the presence of LNAPLs.  However, this
 information may  be useful to  delineate  the presence of LNAPL
 along a vertical profile when limited resources prevent the use
 of more expensive, quantitative techniques.

 Screening of soil headspace  vapors using a flame ionization
detector was also found  useful in evaluating the presence of
volatile NAPL (Cohen et al., 1992). However, a poor
correlation between NAPL saturation in the samples and
                                                         10

-------
headspace vapor concentrations was obtained.  This result
was principally due to physical constraints on contaminant
concentrations in the vapor phase.

Chemical analysis of soil samples will generally be required to
identify individual compounds and their concentrations and aid
in evaluating NAPL present at low residual saturations (Huling
and Weaver, 1991).  Chemical analyses do not directly
distinguish  between gaseous, aqueous, sorbed, and NAPL
contaminants.  However, relatively high constituent
concentrations indicate the constituent may be present as a
NAPL or sorbed to subsurface materials (Huling and Weaver,
1991). An example  of sampling and analysis techniques
applied at the site of an aviation gasoline spill to determine the
LNAPL distribution in the capillary fringe is described by
Ostendorf et al. (1992).

Feenstra et al. (1991) proposed a method to indirectly assess
the presence of NAPL in soil samples by applying equilibrium
partitioning theory. Application of this method to DNAPL
assessments is explained in U.S.EPA (1992b). In the
absence of NAPL, there is a theoretical maximum contaminant
mass which can be present in a sample. The maximum
constituent concentration in the sample is dependent on the
constituent solubility in water, the sorptive capacity of the soil,
and the saturated soil gas concentration. If NAPL is present,
the constituent concentration detected in the sample will
exceed the calculated maximum concentration and the
calculated constituent concentration in pore water would
exceed the effective constituent solubility. This method
requires determination of soil moisture content, organic
carbon content, porosity, potential LNAPL composition,
sorption parameters, effective solubilities, and total constituent
concentrations in soil samples.  The methods are most useful
when relatively large quantities of NAPL are present, pore
water/soil partitioning coefficients are determined rather than
estimated from available literature,  and effective  solubilities
are measured. However, many of these data may be
unavailable at many sites and the methods are less reliable
when parameters are estimated rather than measured (Cohen
and Mercer, 1993; Feenstra et al., 1991).

Cone Penetrometer

The cone penetrometer (ASTM Standard D3441) is a
geotechnical tool that originated in the construction industry.
This tool is  capable of rapidly providing valuable stratigraphic
information useful in assessing potential LNAPL distribution.
Data collected during penetrometer testing have also been
used to evaluate soil saturation, potentiometric surfaces, and
horizontal soil permeability. In its simplest form, truck-
mounted hydraulic rams force a cone-shaped instrument
containing strain gages through the soil, sending back
continuous  readings of tip resistance, sliding friction, and
inclination.  Correlations developed  between subsurface
materials cored in adjacent borings  and corresponding
penetrometer measurements may allow rapid interpretations
of stratigraphy in each test hole (Chiang et al., 1992).
Penetration depths depend on the subsurface material
properties but are often limited to less than approximately
thirty meters.

Modified penetrometers have been equipped to collect
subsurface samples of fluids and aquifer materials useful in
delineating  LNAPL (Chiang et al., 1992).  Other researchers
are incorporating additional in situ sensing technology into
cone penetrometers (Seitz, 1990) for real-time detection of
organic contaminants which may eventually provide a rapid
means for defining LNAPL distribution. It should be noted that
such methods are still in the developmental stage and should
be applied with caution.

Geophysical Methods for Hydrogeologic Characterization

Surface geophysical methods, where applicable, allow non-
invasive investigation of subsurface physical and chemical
properties. Certain techniques, including seismic, electrical,
magnetic, and ground-penetrating radar, have been applied
successfully to site characterization (e.g., Benson et al., 1982;
Walther et al., 1986). These methods may provide valuable
hydrogeologic information concerning lithologic and
stratigraphic boundaries, fracture orientation,  locations of
underground utilities, and depths to ground water and
bedrock. Variations of these and other methods  are useful
borehole techniques for hydrogeologic site characterization
(Keys and MacCary, 1971; Keys, 1989). Use of such
techniques should be considered as part of an integrated
approach to  delineation of hydrogeologic controls on ground-
water flow and potential LNAPL migration.

LNAPL

Light nonaqueous phase liquids may exist as  continuous, free-
phase liquids and/or as residual liquids trapped by capillary
forces above and below the water table. Evaluation of LNAPL
remedial options requires delineation of LNAPL distribution,
composition, and properties. Installation of wells and
excavations  coupled with previously discussed soil sampling
and analysis generally will be required to estimate distribution
and potential mobility.  LNAPL fate, transport, and distribution
has been presented from a theoretical and conceptual model
perspective.  However, site characterization may be
technically challenging due to the heterogeneity of subsurface
media and variability of subsurface conditions. It is not
uncommon to observe a "patchy" distribution of LNAPL over a
relatively small area at a site, or the transient  presence of
LNAPL in a well.

Mobile LNAPL

Monitoring wells, borings, and test pits installed in areas of
potential  LNAPL releases are one of the key sources  of
information concerning the distribution of mobile LNAPL.
Wells used for this purpose must be screened across the air/
LNAPL interface and the water table to provide uninhibited
access for floating LNAPL (API, 1989). The screen generally
should encompass the water table and LNAPL during
seasonal fluctuations in ground-water elevations. Additional
design considerations include use of appropriate construction
materials. In general, conventional construction practices use
a filter pack coarser than the surrounding formation to allow
entry of the LNAPL. However, research regarding
construction  materials specifically designed for LNAPL
monitoring is relatively limited.  Recent studies (Hampton and
Heuvelhorst, 1990; Hampton et al., 1991) of filter packs for
LNAPL recovery wells indicated traditional design practices
may not produce an optimum recovery well. These studies
indicated LNAPL recovery rates were increased  using packs
with a grain size approximately half of conventional
recommendations for recovery wells. Based on such  results,
                                                         11

-------
it appears that additional research regarding optimum filter
pack design for LNAPL monitoring wells may be warranted. In
addition, compatibility of materials, including pumps, casing,
screens, and bentonrte, with the particular LNAPL must be
considered (Mercer and Cohen, 1990). The use of additional
tools such as well points and grab samplers driven to desired
depths (e.g., Smolley and Kappmeyer, 1991) may provide
useful screening information for monitoring network design.

Non aqueous phase liquids held in pore spaces under tension
as a result of capillary forces are not free to migrate to wells
and boreholes. Such contamination may be more extensive
than the mobile LNAPL and  may be a major source for
continuing contamination of ground water.  Monitoring wells
are not useful for defining this contamination (Abdul et al.,
1989). Thus, lack of detection of mobile LNAPL is not an
indication of the absence of such liquids.

Apparent LNAPL Thickness

The LNAPL thickness measured in a monitoring well has been
reported to typically exceed the LNAPL-saturated formation
thickness by a factor estimated to range between
approximately  2 and 10 (Mercer and Cohen, 1990).  Due to
this difference, the LNAPL thickness measured in a monitoring
well has been referred to as an apparent thickness (Figure
10). This difference generally will not be uniform throughout a
site as a result of heterogeneity (Abdul et all,  1989).

In a well screened across the water table and capillary fringe
(Figure  10), the difference has been related to several factors.
These factors include capillary forces in the formation,
hydrocarbon density, and volume/rate of LNAPL release (e.g.,
Blake and Hall, 1984; Hall et al., 1984). In general, the
difference increases with decreasing grain size of the
formation materials due to increased capillary forces and
capillary fringe height (Mercer and Cohen, 1990).  The
ground-water elevation in the well will be lower than the
elevation of the capillary fringe on which the mobile LNAPL
may accumulate as the capillary fringe does not exist within
the well. The LNAPL may then migrate to the well, a  low point
on the capillary fringe. However, thick LNAPL accumulations
in the formation may depress the capillary fringe. In this case,
the difference between the apparent thickness and the true
thickness may be reduced (Testa and Paczkowski, 1989).
This difference also increases with increasing density of the
LNAPL due to  increased depression of the  ILNAPL/water
interface in the well (Hall et al.,  1984).  The weight of the
LNAPL column in the well depresses the interface.  The
greater the LNAPL density, the greater the weight of the
LNAPL column and the greater the depression of the  LNAPL/
water interface.

The relative difference between actual and  apparent LNAPL
thicknesses may also be affected by the existence of  LNAPL
perched on lower permeability layers above the water table
and capillary fringe. The apparent LNAPL thickness
measured in the well may  be related to the  height of the
mobile, perched LNAPL above the water table (Testa and
Paczkowski, 1989). Wells installed through such perched
zones provide potential pathways for LNAPL flow.

Fluctuations in ground-water elevation  may affect the LNAPL
thickness measured in wells (Blake and Hall, 1984;
Kemblowski and  Chiang,  1990;  Yaniga, 1984). A gradual
                                        Well
   Capillary
    Fringe
                                                  Well
                                                 Screen
                                              Apparent
                                              LNAPL
                                             Thickness
                                              Water
                       LNAPL Trapped by
                       Capillary Forces

                       Mobile LNAPL
Figure 10. Simplistic conceptualization of LNAPL thickness
         measured in well and LNAPL distributed information.
decline in the water table may result in increased volume of
mobile LNAPL and increased apparent thickness due to
drainage from the unsaturated zone (API, 1989). Preferential
fluid flow through the well, particularly in low permeability
formations, may also result in increased apparent LNAPL
thickness during a water-table decline (Kemblowski and
Chiang, 1990). In periods of rising water-table, a reduction in
apparent thickness due to compression of the capillary fringe
may occur (Blake and Hall, 1984; Yaniga, 1984). A rising or
falling water table may promote entry of mobile LNAPL into
areas not previously contaminated with these liquids or
regions of lower LNAPL saturation. This results in trapping of
additional LNAPL in soil pores with reduction in the volume of
mobile LNAPL (Figure 8).  Changes in apparent LNAPL
thickness correlated with tidal induced fluctuations in  ground-
water elevation have also been  noted in several studies (e.g.,
Huntetal.,  1989).

Measurement Techniques for Apparent LNAPL Thickness

Common methods for measuring the apparent LNAPL
thickness in wells  include the use of measuring tapes coated
with water-sensitive and hydrocarbon-sensitive pastes,
interface probes,  and transparent bailers (API, 1989).  A steel
measuring tape coated with pastes sensitive to the presence
of water and hydrocarbons may be used to determine the
depths to the  air/LNAPL and LNAPL/water interfaces by
monitoring color changes in the pastes. Alternatively, an
electronic probe designed to detect the electrical
conductivities of different fluids (interface probe) may be used
to determine the depths to the interfaces.  A third common
method is the direct measurement of LNAPL thickness in a
transparent, bottom-filling bailer. The bailer is slowly  towered
until it spans the LNAPL layer, withdrawn, and LNAPL
                                                        12

-------
thickness is measured.  LNAPL thickness in the bailer may be
slightly greater than the apparent thickness present in the well
due to fluid displacement by the bailer walls (API, 1989;
Mercer and Cohen, 1990).  Detection of clear or colorless
LNAPLs may be enhanced by addition of a hydrophobic dye
that preferentially stains the organic liquid (Cohen et al.,
1992).

Recharge and Ball-Down Tests

Methods used to estimate mobile LNAPL thickness in the
formation have included studies referred to as bail-down,
recovery, or recharge tests (e.g., Hughes et al., 1988;
Gruszczenski, 1987). These tests involve monitoring LNAPL
recharge to a well following its removal by pumping or bailing.
The mobile LNAPL thickness in the formation at each well is
then estimated by various interpretations of depth-to-product,
depth-to-water, and product thickness as a function of time.
Several potential sources for error exist in the performance
and interpretation of these tests. In general, these
procedures have not been adequately proven in a variety of
field situations (Testa and Paczkowski, 1989) and do not
provide information concerning LNAPL trapped by capillary
forces (Durnford et al., 1991). However, the performance of
such tests may provide qualitative information concerning the
potential for LNAPL recovery using conventional  pumping
technology (API, 1989).

Apparent LNAPL Thickness Relationships

Studies regarding the relationship between apparent  LNAPL
thickness and in situ LNAPL distribution have been reported
by many researchers (e.g., Ballestero et al., 1994; Blake and
Hall, 1984; Farr et al., 1990; Gruszczenski, 1987; Hall et al.,
1984; Hampton and Miller,  1988; Kemblowski and Chiang,
1990; Lenhard  and Parker, 1990; Pantazidou and Sitar, 1993;
Testa and Paczkowski, 1989; Yaniga, 1984; Yaniga and
Warburton, 1984). Methods proposed for estimating in situ
mobile LNAPL thickness and, ultimately, LNAPL volume  range
from simple relationships based on hydrocarbon density (e.g.,
De Pastrovich et al., 1979) to more complex relationships
incorporating properties of the porous media (e.g., Farr et al.,
1990). Commercial software packages implementing some of
these methods are currently available.

Based on laboratory studies and a review of many of  these
methods, Hampton and Miller (1988)  concluded that the
relationships investigated in their study were not sufficient to
reliably predict  hydrocarbon thickness in the formation. Of
these relationships, the De Pastrovich et al.  (1979) equation
was found to yield crude, order-of-magnitude approximations
of mobile LNAPL thickness. Wagner  et al. (1989) compared
estimates using various techniques including simple and
complex relationships, bail-down tests, and chemical  analysis
of soil samples. The study indicated that estimates from bail-
down  tests, analysis of soil samples from a test pit, a
developmental hydrocarbon-sensing probe,  and the
relationship proposed by De Pastrovich et al. (1979) yielded
comparable results at one field site.  However, none of the
aforementioned methods have been adequately evaluated
under a variety of controlled and field conditions.

Farr et al. (1990) and Lenhard and Parker (1990) developed
methods for evaluating LNAPL volume in porous media under
equilibrium conditions based on fluid and media properties
and apparent LNAPL thickness.  In a controlled study,
Wickramanayake et al. (1991) compared the methods
proposed by De Pastrovich et al. (1979), Hall et al. (1984),
and Lenhard and Parker (1990) to estimate LNAPL volume
from a known release.  A known quantity of JP-4 fuel was
released into a model aquifer equipped with observation wells.
In this study, the method proposed by Lenhard and Parker
(1990) provided the best estimate of LNAPL release after the
system had reached equilibrium. However, all estimates were
within an order of magnitude of the actual release volume.

Durnford et al. (1991) identified several potential limitations in
applying many of these relationships. These problems
include:

     1.  Water table fluctuations can result in differences in
        apparent LNAPL thickness without significant in situ
        changes.  This is a major source of uncertainty in
        estimating in situ distribution.
     2.  Several of these relationships are based on soil and
        fluid  properties that require measurements of
        capillary pressure-saturation curves which are
        difficult to obtain and may not be representative of
        actual field conditions.
     3.  Spatial variability in subsurface properties (i.e.,
        heterogeneity) and the effects on apparent and in
        situ LNAPL thickness are not easily evaluated using
        these relationships.
     4.  LNAPL held by capillary forces above and below the
        water table, which is an important potential source for
        ground-water contamination, is not addressed using
        many of these methods.

In addition, many of the methods are based on an assumed
equilibrium distribution of the fluids.  Such assumptions will not
be applicable at many sites. This implies that simple
relationships and proportionalities may not be sufficient to
estimate mobile LNAPL thickness or volume from apparent
thickness information.  In addition, methods requiring
estimates of subsurface media properties may be subject to
much uncertainty due to heterogeneity and uncertainty in
parameter estimates. However, the techniques described
above may yield order-of-magnitude estimates of mobile
LNAPL distribution at some sites.  In summary, proven field
methods for accurate and  reliable estimation of mobile LNAPL
volume using  well thickness information are not currently
available.  Further research and development of methods for
directly assessing subsurface LNAPL distribution are
warranted.

Geophysical Methods for Contaminant Detection

Application of geophysical methods for direct detection of
organic contamination, including LNAPL, is currently an area
of research and development (U.S.EPA, 1992a). Surface
geophysical methods which have been applied with limited
success at a small number of sites include ground penetrating
radar, complex resistivity, electromagnetic induction, and
direct current  resistivity (e.g., Olhoeft, 1986; King and Olhoeft,
1989; Holzer, 1976). Borehole techniques which have been
studied include dielectric logging (Keech, 1988) and a driven
probe equipped with galvanic and conductivity circuits for
sensing  soil moisture and a hydrocarbon indicator for sensing
LNAPL (Hampton et al., 1990).  Such techniques are currently
in the developmental stage and  should generally be
                                                        13

-------
 considered research applications.  The utility of geophysics at
 most sites will not be in the direct detection of LNAPL, but in
 hydrogeologic site characterization (U.S.EPA, 1992a).

 LNAPL Samples

 Determination of physical and chemical properties of LNAPL
 obtained from wells or separated from soil samples will often
 be required to evaluate many aspects of LNAPL site
 characterization and remedial design. For example,
 information concerning physical properties such as density
 and viscosity may be used to assess LNAPL mobility and
 distribution (Reidy et al., 1990; Cohen and Mercer, 1993) and
 potential extraction designs.  Analyses of LNAPL will be
 necessary to determine the chemical composition which may
 be used to compute the effective solubility of LNAPL
 components, identify potential LNAPL sources, and evaluate
 applicability of certain remedial technologies such as soil
 vapor extraction.

 Depending on the study objectives and potential LNAPL
 composition,  many analytical methodologies may be
 applicable for chemical analysis of LNAPL samples. Common
 techniques include infrared spectrometry, gas and  liquid
 chromatography, mass spectrometry, and nuclear magnetic
 resonance spectrometry.  Such techniques are used to
 qualitatively and quantitatively determine hydrocarbon
 composition.  Cohen and Mercer (1993) describe many of
 these techniques in the context of characterizing DNAPL
 samples.  This information is  also applicable to LNAPL sample
 characterization. Multiple techniques may be required to
 characterize complex, multicomponent LNAPL mixtures.  In
 addition,  analyses for indicator parameters such as total
 petroleum hydrocarbons and  total organic carbon may be
 useful in screening level investigations at some sites.

 Current guidance dictates LNAPL samples should be obtained
 prior to purging of the well. Purging of the LNAPL from the
 well prior to sampling may be conceptually desirable to obtain
 a more representative sample.  However,  in many cases
 purging may result in an inability to sample due to slow or no
 LNAPL recharge and em unification of the sample.  Specific
 sampling techniques should be evaluated with regard to study
 objectives and site conditions. A bailer will generally be
 adequate for LNAPL sample collection from wells (API, 1989).
 Additional equipment such as a bladder pump, peristaltic
 pump, or specialized equipment  used for LNAPL recovery
 may be useful at many sites.

 So/7 Gas

 Depending on site conditions, soil gas analysis for volatile
 organic compounds may be useful  in locating contaminants
 present in the subsurface. Several techniques have been
 developed for conducting such surveys (e.g., Marrin and
 Kerfoot, 1988; Thompson and Marrin, 1987).  However,
 limitations in the use and interpretation of data from soil gas
 surveys exist (Kerfoot, 1988; Marrin, 1988). Soil gas surveys
 may provide qualitative information useful as a screening tool
 in delineating  areas of LNAPL contamination as well as
 aqueous-phase contamination (e.g., Evans and Thompson,
 1986; Devitt et al., 1987). Anomalously high soil  gas
concentrations may be an  indication of NAPL in the vadose
zone near the sample location.
 Ground Water

 Ground-water elevations in wells containing LNAPL require
 correction for the depression of the LNAPL/water interface in
 the well to obtain total hydraulic head.  The depression is
 caused by the weight of the hydrocarbon.  The correction is
 accomplished by multiplying the apparent LNAPL thickness in
 the well by the specific gravity of the LNAPL  The result is
 then added to the elevation of the LNAPL/water interface to
 obtain the total hydraulic head (API, 1989; Hudak et al., 1993).
 The computed hydraulic head facilitates a more accurate
 assessment of hydraulic gradient which leads to better
 conclusions regarding potential contaminant sources, extent
 of free-product accumulation, and optimal areas for focusing
 remediation efforts (Hudak et al., 1993). It should  be noted
 that this approach may be inappropriate for conditions in
 which LNAPL  enters the well from a low-permeability unit
 perched above the water table.

 The influence of this correction on data interpretation is
 dependent on  several factors, including apparent LNAPL
 thickness and  hydraulic gradient. The greater the thickness of
 LNAPL in the well and the lower the hydraulic gradient, the
 greater the potential influence of this correction. The
 necessity of performing this calculation should be evaluated
 on  a site-specific basis.  Accumulation of LNAPL in a well
 introduces an additional degree of uncertainty into the
 calculation of hydraulic head and gradient. Density of the
 LNAPL obtained from the well must be  measured and, often,
 will not be constant  across a site due to such factors as
 LNAPL release history and differences  in the degree of
 "weathering" following release.  Measurement of the apparent
 thickness using any of the methods previously described is
 subject to a higher degree of uncertainty than measurement of
 ground-water elevation  in a well which does not contain
 LNAPL. The additional uncertainty introduced by LNAPL
 accumulation in wells is site specific and should be evaluated
 during any investigation.

 Although contaminant concentrations in ground-water
 samples may not approach the effective solubility of each
 compound, analyses for LNAPL components may aid in
 evaluating the  potential presence of LNAPL and delineating
 potential source areas.  Several factors may account for
 lower-than-expected concentrations of constituents in ground
 water from LNAPL zones. First, the effective solubility of a
 single chemical from a multicomponent LNAPL is less than the
 solubility of the pure chemical in water.  The difference is often
 an order of magnitude or more. Other explanations for the
 relatively low concentrations of dissolved constituents in
 LNAPL zones may include variability in  subsurface LNAPL
 distribution, mixing of ground water from different intervals in a
 well during sampling, and effects resulting from non-uniform
 ground-water flow (Cohen etal.,  1992).  Non-equilibrium
 dissolution may also be a factor at some sites (Powers et al
 1991).

 REMEDIATION

Applicability of  potential remedial technologies depends on
site-specific hydrogeologic characteristics, nature/distribution
of contaminants,  and remedial objectives. Technologies for
removal of mobile LNAPL exist and may be applicable at
some sites. However, technologies for removal of residual
LNAPL, particularly those liquids trapped in the saturated
zone, are not well developed. Subsurface restoration to pre-
                                                        14

-------
contamination conditions may require removal of virtually all
LNAPL and much of the contamination sorbed to aquifer
material.  Technological limitations to complete LNAPL
removal may exist at many sites.

Excavation

The remedial alternative(s) selected for an LNAPL site may
depend on the volume of LNAPL released. If the volume is
small enough to be retained within the upper portion of the
vadose zone, excavation and above-ground treatment/
disposal may be selected in lieu of in situ remediation. Some
of the advantages of excavation are:

     •   Costs of shallow excavation are generally low
        compared to other remediation technologies.
     •   Remediation time is relatively short.
     •   Experienced contractors are more readily available.
     •   A wide variety of contaminants often can  be
        remediated with a high degree  of reliability.

The potential disadvantages include:

     •  Contamination may not be completely removed due
        to  inability to delineate LNAPL distribution in detail.
     •  Structures, roads, or other features  can restrict the
        area available for excavation.
     •  Air emissions of volatile constituents may be
        significant.
     •  Costs of deep excavation may  be high.
Mobile LNAPL Recovery - Trenches/Drains/Wells

Recovery potential for mobile LNAPL is controlled by such
factors as LNAPL viscosity and density and relative
permeability (Testa and Paczkowski, 1989).  Liquids trapped
by capillary forces are not recoverable using conventional
trench, drain, or well systems.  High LNAPL viscosity, high
residual water saturation, and low permeability reduce LNAPL
recovery rates.  Under optimum circumstances, these systems
may remove less than 50% of the total LNAPL volume in the
subsurface (Abdul,  1992). It is estimated that only 20% to
30% of the total release volume is typically recovered (Testa
and Paczkowski, 1989). The remaining LNAPL will generally
be sufficient to result in continued ground-water
contamination.  However, there may be important benefits to
mobile LNAPL recovery other than simple contaminant mass
removal.

Some of the other potential reasons for recovery  of these
liquids include:

     •  Mobility reduction.  Residual LNAPL held by capillary
        forces is relatively immobile under normal
        hydrogeologic conditions.  Recovery of mobile
        LNAPL will limit the migration of these liquids.

     •  Increase LNAPL transformation.  Reduction in
        LNAPL saturation  may increase LNAPL surface area
        and may result in more rapid dissolution,
        degradation, and volatilization of the LNAPL. This
        may be an advantage to the remediation program at
        some sites.
                                                                             Recovery Well with
                                                                               Open Interval
                                  Plan View
           Cross Section
                                                                                    After API (1989)
                        Figure 11. Plan and cross sectional views of a drain designed for mobile LNAPL recovery.
                                                        15

-------
 Trench/Drain Systems

 Trenches/drains may be used to recover mobile LNAPL at
 shallow depths. Generally, they are used at sites where the
 LNAPL is located within 15 to 20 feet of the surface with
 access for construction equipment (U.S.EPA, 1988).
 Depending on available excavation equipment, costs, and soil
 stability, deeper drains may be cost effective (API,  1989).
 Trench/drain systems provide the  most hydraulically efficient
 means for removing fluids from the aquifer.  Therefore, these
 systems often are well suited for low-permeability units and
 heterogeneous sites which would  require a large number of
 wells to control LNAPL flow (API, 1989).  Trenches are usually
 excavated perpendicular to the general direction of ground-
 water flow, downgradient and/or within mobile LNAPL (Figure
 11). The trench generally should be as long as  the width  of
 the mobile LNAPL to provide containment. It should be
 excavated to a depth several feet below the  lowest expected
 seasonal water-table elevation to prevent the migration of
 LNAPL below the collection system.  However, trenches for
 recovery of LNAPL perched on low permeability layers should
 not penetrate the supporting unit allowing further downward
 migration to the water table. In general, trench width is not a
 factor in LNAPL recovery (U.S.EPA,  1988) and may be as
 narrow as possible to reduce costs (API, 1989).  A low
 permeability barrier may be installed on the downgradient side
 of the trench to limit hydrocarbon migration while allowing
 water to flow underneath.  However,  such  a barrier may not be
 an essential component of the design.  In case of a trench/
 drain constructed in fine-grained native materials using a
 coarse  grained fill, a capillary barrier effect may exist. Even
 without installation of an artificial barrier, the non-wetting
 LNAPL may tend to be trapped in the trench due to the higher
 elevation of the capillary fringe in the formation on each side
 of the trench. As long as product is removed from the trench,
 LNAPL migration beyond the trench may be limited.

 Open trenches may be converted to drains by backfilling with
 appropriate graded filter materials (e.g., gravel).  Sumps or
 wells may be installed along the  drain to collect LNAPL (API,
 1989).  Perforated pipe installed in the trench and connected
 to the sumps may be used to improve system efficiency.
 Recovery of LNAPL from trenches/drains or wells may be
 accomplished using several options including pumps for
 collecting total fluids (i.e., LNAPL and water) and skimming
 systems and pumps for collecting only LNAPL (API  1989-
 U.S.EPA, 1988).

 Trenches/drains may be operated to recover only LNAPL
 migrating under the influence of the natural hydraulic gradient.
 A system operated in this manner would generally result in a
 relatively low recovery rate. Recovery time frames may also
 be relatively long unless multiple trenches/drains are used.
 Such systems should generally use continuous LNAPL
 recovery to prevent accumulation and migration around the
 ends of  the trench (API, 1989). Ground-water extraction to
 increase hydraulic gradients toward the trench/drain may be
 used to  increase LNAPL recovery rate and establish
 hydrodynamic control.  However, depressing the water table
 may also result in LNAPL migration into deeper portions of  the
 saturated zone previously uncontaminated  by these liquids. In
this situation, residual LNAPL will remain trapped below the
water table following mobile LNAPL recovery. The potential
 consequences of such system operation should be carefully
 considered before implementation.

 Recovery Wells

 Wells used to recover mobile LNAPL offer more flexibility in
 system design and operation than trench/drain systems.
 Wells may be most useful at sites with moderate to high
 hydraulic conductivities, but may be used in less conductive
 materials. Conventional recovery well designs have relied on
 modified water-well design procedures to maximize well
 efficiency and increase LNAPL recovery rates (Blake and
 Lewis, 1983). However, well construction materials and
 recovery equipment compatible with the LNAPL should be
 used (Mercer and Cohen, 1990). Potential  fire and explosion
 hazards should also be considered.

 Wells may be designed to remove only LNAPL, LNAPL and
 water separately, or a combined fluid mixture (API, 1989;
 Blake and Lewis, 1983; U.S.EPA, 1988). Well construction
 depends on system design. Screens generally are set across
 the air/LNAPL/water interfaces in the well.  Such  screens
 should be long enough to encompass the anticipated changes
 in position of these interfaces due to pumping and other
 influences.  Maximization of open screen area through use of
 continuous wire-wrapped designs has been recommended
 (Blake and Lewis, 1983) to maximize well efficiency and
 extend intervals between maintenance.  Well diameter
 depends,  in part, on the proposed pumping  equipment.  Many
 conventional designs have used wells of 6-inch diameter or
 larger. However, recent equipment innovations have allowed
 use of smaller diameter wells.

 Standard water-well design procedures traditionally have been
 used  for gravel pack design.  However, studies by Sullivan et
 al. (1988) indicate that filter packs which are too coarse may
 reduce hydrocarbon recovery efficiency.  Laboratory studies
 reported by Hampton and Heuvelhorst (1990) indicate that
 LNAPL recovery may be enhanced using hydrophobic filter
 packs and a grain size approximately half as large as
 traditional designs. Further laboratory studies (Hampton et al,.
 1991) again indicated that hydrophobic materials  were
 superior to other materials for filter pack construction. Of the
 remaining materials which were tested, the mixture containing
 the least quartz, most angular grain shape, and least uniform
 size distribution produced the most efficient  filter pack.
 Further research, including field studies, regarding well
 construction for LNAPL recovery appear to be warranted.

 Fluid recovery rate for wells designed to recover only LNAPL
 generally will be relatively low. Thus, the reduction in
 hydraulic head due to removal may be minimal. This results in
 a relatively small capture zone around each  well and,
generally,  hydrodynamic control is not maintained. Conditions
 in which such systems may be most applicable include
situations where water-table depression is not necessary or
desired and water treatment/disposal capacities are limited
(API, 1989).

Equipment designed for recovery of only LNAPL operates
using  a variety of mechanisms (API, 1989) including:
                                                        16

-------
     1.  top mounted intakes allowing fluid collection from the
        LNAPL/water interface,
     2.  density-sensitive float valves,
     3.  conductivity sensors for pump activation, and
     4.  hydrophobia filters that preferentially allow
        hydrocarbons to pass.

Such equipment can recover LNAPL to a thickness of a
fraction of an inch for systems equipped with conductivity
sensors and to a sheen for systems using a hydrophobic
membrane (U.S.EPA, 1988).

Wells designed to recover total fluids (i.e., LNAPL and water)
may use surface  mounted suction-lift or submersible pumps
(API, 1989). Recent applications of this technology are similar
to wellpoint dewatering systems used in the construction
industry (Hayes et al., 1989).  In this application, several
closely spaced driven or drilled wells may be manifolded to a
single high volume vacuum pump. However,  practical
recovery depths using suction-lift pumps are limited to
approximately 20 feet. Total fluids recovery systems generally
require LNAPL/water separation which may be difficult due to
emulsification.  Emulsification problems can be minimized by
proper pump selection. Centrifugal pumps generally will
emulsify the LNAPL/water mixture more than surface
diaphragm pumps or down-hole pneumatic pumps (API,
1989). Aqueous-phase constituent concentrations in
recovered water are also relatively high due to LNAPL/water
mixing.  Although relatively high yields are possible (Hayes et
al., 1989), total fluids removal may be effectively used in
situations where the hydraulic conductivity is too low to  permit
the efficient use of higher capacity pumps (API, 1989; Blake
and Lewis, 1983).

Operation of recovery well systems using total fluids recovery
may sometimes be enhanced by sealing the wellhead and
creating a vacuum within the well (API, 1989; Hayes et  al.,
1989). Vacuum generation enhances recovery by reducing
pressure within the casing, increasing the effective head
difference between the formation and the well.  Wells
constructed using this design may be installed with the top  of
the screen below the mobile LNAPL (Hayes et al., 1989). The
water table is depressed sufficiently during  pumping to allow
LNAPL to enter the well.  This results in less vacuum loss
through minimal screening of the unsaturated zone.

Dual pump systems (Figure 12) use one pump located near
the bottom of the well to extract only water, creating a cone of
depression to initiate LNAPL movement towards the well (API,
1989; Blake and Lewis, 1983; U.S.EPA,  1988). A separate
                               Well
                              Screen
                                                                            Water
                                                                            Pump
                                                                                After API (1989)
                  Figure 12. Conceptual system design for separate recovery of LNAPL and water in a single well.
                                                        17

-------
 pump located in the upper portion of the screened interval is
 used to remove LNAPL. Both pumps rely on sensors to
 ensure that only water and only LNAPL are extracted in their
 respective pumps (API, 1989). Dual pump systems may also
 use submersible or surface mounted pumps. Positioning of
 the water pump generally should  be well below the LNAPL/
 water interface  to minimize constituent concentrations and
 treatment requirements for recovered ground water.  Single
 units, combining separate LNAPL and water pumps, equipped
 with sensors to prevent pumping mixed fluids are also
 available (U.S.EPA, 1988).

 Dual well LNAPL recovery systems extract water and LNAPL
 from separate wells. As with single well/dual pump systems,
 one well produces only water and induces a hydraulic gradient
 to promote LNAPL movement. The adjacent well is used to
 recover LNAPL that accumulates in the cone of depression
 Conditions under which dual pump and dual well systems are
 normally used include (API, 1989):

      1.  Water-table depression to increase LNAPL recovery
         rates is desirable and not detrimental to the
         remediation objectives.
      2.  Hydraulic conductivity and saturated thickness are
         large enough to sustain flows of separate fluid
         streams.
      3.  LNAPL/water separation and treatment capacities
         are limited.

 Installation of horizontal wells is an emerging technology in
 the environmental field (Morgan, 1992).  Such wells  could
 greatly increase contaminant removal efficiency, including
 LNAPL removal, at some sites. Reported use of this
 technology at contamination sites (e.g., Looney et al., 1992;
 Oakley et al., 1992) has been limited. Additional information
 regarding construction and use of horizontal wells is available
 in U.S.EPA (1994).

 System Design and Operation

 At sites where a systematic design approach has been used,
 traditional LNAPL extraction system designs generally have
 relied on estimation of the recovery well capture zone. Well
 locations are then selected to capture and remove
 hydrocarbons (e.g., API, 1989; Blake and Lewis, 1983).
 Injection of recovered ground water also has been used to
 increase  LNAPL recovery rates by increasing hydraulic
 gradients and provide a hydraulic barrier to LNAPL migration
 (e.g., Zinner et al., 1991).  Careful evaluation of the effects of
 reinjecting treated/untreated ground water is recommended to
 minimize the potential for increasing subsurface
 contamination.

 Much of the optimization of LNAPL  recovery has been
 conducted in the field through manipulation of LNAPL and
 water pumping rates. Several factors should be considered
 during system design and operation. One important  design
 consideration is  the effect of water-table depression used to
 increase LNAPL recovery rates (API, 1989; Blake and Lewis,
 1983; Chiang et al., 1990). Although high water extraction
 rates initially may yield a higher hydrocarbon recovery rate,
the ultimate recovery may be greatly reduced.  Much of the
 LNAPL that was originally mobile may be smeared into
 uncontaminated portions of the aquifer as the water table is
 depressed, trapping it as immobile, residual LNAPL in soil
 pores.  This residual LNAPL will remain trapped below the
 water table when the recovery system is turned off providing a
 continuing source for ground-water contamination. At many
 sites, multiple wells pumping water and LNAPL at lower rates
 may ultimately recover more LNAPL than fewer wells pumping
 water at higher rates to create large water-table depressions
 (API, 1989).  In addition, if the recovery rate exceeds the
 LNAPL migration rate to the well, water saturation around the
 well may increase, resulting in lower relative permeability with
 respect to LNAPL and reduced recovery (Abdul, 1988; Chiang
 etal., 1990).

 Similar considerations exist for systems using water injection
 to increase LNAPL recovery rates. Water-table mounding
 may result in pushing LNAPL upward into previously
 uncontaminated  intervals (Testa et al.,  1992).  Injection may
 also cause undesired lateral LNAPL migration. These
 potential effects should be considered during system design.

 Additional variables which may affect optimal LNAPL recovery
 rates include thickness of the mobile LNAPL and formation
 permeability to the LNAPL (Charbeneau et al., 1989; Chiang
 et al., 1990). As the mobile LNAPL thickness  increases, the
 transmissivity of the formation with respect to LNAPL also
 increases resulting in increased recovery.  In low permeability
 situations, the overall system recovery rate may become a
 linear function of the number of wells.  This is often due, in
 part, to a lack of hydraulic interference between wells.

 In recent studies, several researchers have recommended
 strategies for maximizing the overall recovery of LNAPL from
 systems using LNAPL and water extraction. One
 recommended strategy is to pump the LNAPL layer at
 relatively low rates in order to maintain the LNAPL as a
 flowing continuous mass (Abdul, 1988;  Charbeneau et. al.,
 1989; Chiang et al., 1990). Water pumping, if  used, is
 carefully controlled to minimize smearing of the LNAPL layer
 and to prevent upconing of the water table. As expected,
 these studies indicate that this approach results in a relatively
 small capture zone for each pumping well.  Although this
 implies that the total LNAPL recovery rate may be a linear
 function of the number of wells which are used, a greater
 volume of LNAPL may be recoverable than is possible using
 other approaches such as maximizing water-table depression.

 Based on mathematical simulations, Chiang et al. (1990)
 produced a series of nomographs to aid in estimating the
 optimal LNAPL recovery rate for a well under various
 conditions of hydraulic conductivity, LNAPL thickness, density,
 and viscosity. However,  assumptions such as homogeneity of
 subsurface materials and LNAPL thickness were used in
 development of the nomographs, limiting their  applicability at
 many sites. Comparison with site-specific data would be
 required to verify the utility of these nomographs for a specific
 site. Charbeneau et al. (1992) also proposed a general
 method for selecting the optimum  hydrocarbon pumping rate
 in dual water/LNAPL pumping systems as a function of water
 pumping rate. However, this method only aids in estimating
 an LNAPL pumping rate and does not indicate  an optimal
 pumping rate for water.

The time frame required for mobile LNAPL recovery using
extraction wells, trenches, or drains depends on many factors
                                                        18

-------
including LNAPL and porous media properties, LNAPL volume
and distribution, and recovery system design. Time frames
generally will be difficult to estimate due to heterogeneity and
uncertainty in the values of pertinent parameters.  At many
sites the time frame required for recovery will be a direct
function of the number of wells, trenches, and drains installed.

LNAPL Modeling

Several multiphase flow models have been developed which
are capable of simulating LNAPL transport (e.g., API, 1988;
Charbeneau et al., 1989; Faust et al., 1989; Huyakorn et al.,
1992; Kaluarachchi and Parker, 1989; Kaluarachchi et al.,
1990; Katyal et al., 1991; van der Heijde and Elnawawy, 1993;
Weaver et al., 1994). Such models have been used in site
characterization to simulate potential contaminant distribution
and  LNAPL recovery system design. However, multiphase
flow is a complex problem, particularly in a heterogeneous
environment. Models incorporate simplifying assumptions to
facilitate utility. Recognition of the underlying assumptions
and evaluation of the site-specific applicability of the model is
required. Data requirements may also be extensive.  Certain
parameters may not be readily measured in the field due to
site characterization technology limitations.  Many of the
models are sensitive to parameters such as permeability,
porosity, and LNAPL spill history that are often unknown or
poorly defined. Thus, significant uncertainty in the accuracy of
the results may exist, even at relatively well  characterized
sites.

The applicability of these models at many sites may be  limited.
Objectives of modeling, required quality/quantity of
characterization data, and required confidence in the model
results should be evaluated prior to initiation of the exercise
(Huling and Weaver, 1991).  Use of models  as screening level
and site characterization tools may be beneficial at some
sites.  For example, Weaver et al. (1994) developed a simple
collection of integrated models called the Hydrocarbon Spill
Screening Model to help predict the environmental impact on
ground water from LNAPL spills. Included are models for
simulating the downward migration of LNAPL through the
unsaturated zone, radial transport of the oil lens on the water
table, and advection and dispersion of the dissolution products
in ground water.  The modeling package is intended to be
used as a screening tool for estimating the effects of
parameters such as LNAPL loadings, adsorption,  ground-
water velocity,  and other major factors.  Results of simulations
using other appropriate models may also aid in designing
LNAPL recovery systems. However, as in any simulation of
subsurface processes, significant  uncertainty in the accuracy
of the results generally exists and should be considered in the
modeling and remedial design processes.

So/7 Vapor Extraction

Soil vapor extraction (SVE) is a rapidly developing technology
with applications for removal of volatile contaminants from the
vadose zone. In a simple system, air is drawn through the
affected area by applying a vacuum to vapor extraction  wells
(Figure 13), stripping volatile organic compounds into the
moving air stream.  The air containing the organic vapors is
then treated, if necessary, and discharged to the atmosphere.
More complex systems may incorporate trenches, air injection
wells, and surface seals to direct the air flow through the
                               After Huling and Weaver (1991)
Figure 13.  Conceptual design of soil vapor extraction system
          with vapor treatment prior to discharge.
desired remediation zone (e.g., Johnson et al., 1990b; Lyman
and Noonan, 1990; Pedersen and Curtis, 1991).

Success of the SVE system at LNAPL removal is based on
three factors:  1) chemical composition of the LNAPL, 2) vapor
flow rates through the unsaturated zone, and 3) flow path of
the carrier vapors relative to the zone of contamination
(Johnson et al., 1990a).  In general, SVE may be best suited
for the removal of compounds with vapor pressures greater
than approximately 14 mm Hg at 20° C (Sims, 1990). With
multiple component LNAPLs, preferential removal is observed
for chemical compounds with higher vapor pressures and
higher mole fractions in the LNAPL mixture.

Subsurface media must be permeable enough to allow
adequate vapor movement. Vapor flow rates are dependent
on the vacuum drawn in the SVE wells and the properties of
the unsaturated zone (e.g., soil moisture content, soil texture,
macropores, and LNAPL distribution).  The ability to direct flow
through the contaminated zone is another primary concern.
Vapor flow paths to wells are affected  by the degree of  short-
circuiting from the surface that occurs, locations of subsurface
heterogeneities, and the presence of preferential flow paths.
A design process  based on vapor pressure data for the
contaminants and soil/air permeability data was developed by
Johnson et al. (1990a, 1990b) to determine  if a site is
amenable to SVE and to estimate the size of the required
system. Additional information concerning design and
application of SVE systems is available from DiGiulio (1992),
Pedersen and Curtis (1991), and U.S.EPA (1992c).

Ground-water extraction may be used with SVE systems to
counteract upconing of the water table near the SVE well
caused by pressure reduction (Johnson et al., 1990a).
Increased ground-water extraction may dewater deeper
intervals exposing contaminated materials for recovery  by
                                                        19

-------
SVE at some sites. Controlled studies conducted in a model
aquifer by Johnson et al. (1992) indicate this application of
SVE may be very effective in removing much of the residual
NAPL previously trapped below the water table.

Enhancements which have been applied to SVE systems
include in situ soil heating and induced fracturing.
Contaminant volatility, which can be one of the limitations to
the effectiveness of SVE, increases with increasing
temperature. Steam or hot air injection  has been investigated
for use in increasing removal of volatile  and semivolatile
constituents (DePaoli and Hutzler,  1992; Sittler et al.. 1992;
U.S.EPA, 1992a). Studies using radio-frequency heating
have also been performed (U.S.EPA, 1992a).  Field scale
trials have been relatively limited to date.  Hydraulic and
pneumatic methods for  inducing fractures in fine grained
vadose zone materials have also resulted in increased vapor
yields and contaminant  removal during field trials (U.S.EPA,
1993a;U.S.EPA, 1993b).

Another recent enhancement to the SVE concept is the use of
induced air flow to increase biodegradation of contaminants in
the vadose zone  (e.g., Kampbell, 1992; van Eyk, 1992).  The
term "bioventing" has been applied to such systems.  The
system is operated  to deliver oxygen to the indigenous
microbes, thereby promoting degradation of contaminants.
Research efforts  are currently focusing on design issues such
as the need for nutrient  addition and the kinetics of the
biological reactions.
There are practical limitations to the effectiveness and
efficiency of SVE at many sites.  For example, site conditions
resulting in vapor flow around but not through contaminated
zones will result in contaminant mass transfer limitations
(Johnson et al., 1990a; Travis and Macinnis, 1992).
Situations in which such conditions may occur include;
1) contaminated low permeability materials surrounded or
layered with higher permeability materials, and 2) LNAPL
removal from  a relatively thick layer with a high liquid
saturation and a correspondingly low vapor permeability.
Contaminant mass transfer in these cases is limited by
diffusion from the contaminated zone.  The effectiveness of
SVE in removing contaminants from the water-saturated zone
is similarly limited by diffusive transfer.  In general, soil vapor
extraction may be applicable for removal of much of the
LNAPL located above the water table, including liquids
retained by capillary forces, at many sites. Selection of SVE
or conventional pumping technology for removal of mobile
LNAPL will depend on site conditions and remedial objectives.

Air Sparging

Air sparging is a relatively new technology that is being
implemented to remove volatile contaminants below the water
table in unconfined aquifers (e.g., Loden,  1992; Marley et al.,
1992a; U.S.EPA, 1992a). Air is injected from a well screened
in the saturated zone (Figure 14). Two potential objectives
cited for such  systems are: 1) strip volatile hydrocarbons from
the aqueous phase and from any NAPLs present along the
                                              Air
                              Air
                            Channels
Vapors
  4
       Air
                                 Figure 14. Conceptual design of air sparging system.
                                                        20

-------
path of air flow, and 2) add oxygen to the water to encourage
biodegradation of amenable constituents.   The compounds
must be volatile to be stripped and transported to the surface.
Additionally, the compounds must be biodegradable for the
biodegradation component of this technology to be effective.
In the case of multicomponent LNAPLs, a certain fraction of
the LNAPL may be volatile and biodegradable.  However, a
certain fraction of the LNAPL may be relatively  unaffected by
air sparging. After the air make its way to the unsaturated
zone, a soil vapor extraction system is used to remove the
vapors for treatment prior to release to the atmosphere.

Air sparging generally is proposed in conjunction with vacuum
extraction, therefore, design limitations, and concerns
associated  with vacuum extraction are also applicable in air
sparging systems. Specific design parameters  for air sparging
are numerous and detailed discussions are beyond the scope
of this review. Generally, these parameters fall under three
categories: hydrogeological constraints, contaminant
properties and distribution, and system operations (Marley et
al., 1992a;  Brown, 1994). However, the primary design
consideration is contaminant volatility, (i.e., volatilization from
the water phase (Henry's Law) or from the NAPL (Raoult's
Law)).

One limitation of air sparging is the vulnerability to preferential
pathways and heterogeneities.  Short circuiting along
manmade and natural pathways (sheet piles, wells, sparge
points, root holes, high hydraulic conductivity layers,
stratigraphic windows, etc.) reduces the overall effectiveness
of air sparging.  Subsurface heterogeneity results in
channeling air flow through preferential pathways (Ji et al.,
1993; Johnson et al., 1993).  Sparging below an impervious
stratigraphic unit may direct air laterally, spreading
contamination (Brown, 1994). As with other technologies,
overall effectiveness may be reduced due to rate limited
mechanisms such as contaminant desorption and slow
contaminant diffusion from low permeability materials.
Consequently, the radius of  influence of a sparge well is
difficult to evaluate, and may be relatively small.

Additional potential concerns with use of this technology
include increased contaminant migration and in situ
precipitation of dissolved minerals (Marley et al., 19926).
Mechanisms such as formation of gas pockets  may result in
lateral displacement of contaminated ground water or LNAPL
into previously clean areas.  Uncontrolled vapor migration in
the vadose zone may also spread contamination. Clearly,
control of both the vapor emissions in the unsaturated zone
and the ground-water plume generally must be monitored and
maintained. Mineral precipitation due to changes in
geochemical conditions may result in undesired permeability
reduction at some sites.  Adequate site characterization,
system design, and monitoring would be required to mitigate
such concerns.

Controlled  studies of the effectiveness of  this technology are
relatively limited. In a study conducted in a model aquifer
(Johnson et al., 1992) contaminant mass  removal using air
sparging was observed.  However, significant contaminant
mass appeared to remain following sparging. Preferential
pathways and a limited radius of influence were also
observed.  The overall performance of this technology has not
been adequately assessed under a variety of field conditions.
Carefully monitored field demonstrations are required to better
determine applicability and effectiveness.
Enhanced Oil Recovery Technologies

Methods for enhancing oil recovery which were pioneered by
the petroleum industry are being developed for environmental
applications. Technologies under investigation include
injection of hot water or steam, cosolvents (e.g., ethanol),
surfactants, alkaline agents, and polymers. Currently many
researchers are evaluating the applicability of enhanced oil
recovery (EOR) technology to the remediation of LNAPL sites.
Most of these technologies are still experimental. Very little
information regarding field applications is available. Current
reviews of the development of EOR technologies for DNAPL
removal (U.S.EPA, 1992a) and other hazardous waste site
remediations (Sims, 1990) are available.  Information from
these sources is also applicable to LNAPL site remediation.

Primary LNAPL recovery systems (e.g., drains, pumping
wells) will generally result in the removal of significantly less
than 50% of the total LNAPL volume.  Enhanced oil recovery
may remove more of the LNAPL. However, there are practical
limitations to the effectiveness of these techniques.
Conditions such as subsurface heterogeneity, low
permeability units, and relative permeability reductions caused
by the presence of NAPL can prevent remediation fluids from
making contact with the NAPL.  Significant quantities of
LNAPL may remain  in place following EOR applications
(Mercer and Cohen, 1990).

Recovery of LNAPL may be enhanced by injecting fluids to
increase hydraulic gradients, reduce the NAPL/water
interfacial tension, reduce NAPL viscosity, increase wetting-
phase viscosity, and/or increase NAPL solubility (U.S.EPA,
1992a). The application of heat to viscous LNAPL will
enhance mobility by decreasing viscosity and may result in
increased solubility. Volatile NAPL components may also
volatilize and condense in advance of a steam or hot water
flood increasing NAPL saturation and relative permeability.
Steam/hot water/hot air flooding, electrical heating, radio
frequency heating, and conduction heating are possible
techniques which have been investigated (e.g., Davis and
Lien, 1993; Fulton et al., 1991; Hunt et al., 1988a;  Hunt et al.,
1988b; Johnson and Leuschner, 1992; Sims, 1990; Udell,
1992; U.S.EPA, 1992a). Thermal techniques have potential
application for increased removal of LNAPL in both the
saturated and unsaturated zones.  These techniques have
been evaluated in a relatively limited number of field scale
studies (U.S.EPA, 1992a).

Chemically enhanced recovery techniques appear to be
promising technologies for increasing LNAPL solubility and
mobility (U.S.EPA, 1992a).  Injection of surfactant  solutions
has been investigated for use in increasing the solubility of
NAPL constituents and increasing  NAPL mobility through
reduction of interfacial tension (Fountain, 1992). Surfactants
are potentially capable of increasing NAPL solubility by
several orders of magnitude. Injection of alkaline agents has
been  used in the petroleum industry to increase NAPL mobility
through in situ surfactant production resulting from reaction
with organic acids in the NAPL.  Polymer flooding has been
used as a component of water flooding technology by the
petroleum industry.  Polymers are  used to displace some
portion of the residual NAPL by increasing the viscosity of the
water flood. Application of these technologies in the
environmental field has been very  limited (U.S.EPA, 1992a).
In addition, use of cosolvents has been proposed for
increasing  solubility of NAPL compounds (e.g., Augustijin et
                                                         21

-------
 al., 1992; Boyd and Farley, 1992; Rao et al., 1991; U.S.EPA,
 1992a). In general, chemically enhanced recovery of
 hydrocarbons for environmental applications is in the
 developmental stage (U.S.EPA, 1992a).

 Bioremediation

 Practical biological degradation of LNAPL pools has not been
 demonstrated. Bioremediation of immiscible hydrocarbon is
 limited due to the following: (1) NAPLs present a highly hostile
 environment to the survival of most soil microbes,  (2) the basic
 requirements for microbial proliferation (nutrients, terminal
 electron acceptor, pH, moisture, osmotic potential, etc.) are
 difficult if not impossible to deliver or maintain in the NAPL
 (Huling and Weaver, 1991). Correspondingly, bioremediation
 may be limited to the periphery of the NAPL zone in both
 saturated and unsaturated media. It has been postulated that
 biologically-produced surfactants resulting from microbial
 activity near a NAPL have increased the rate of NAPL
 solubilization (Wilson and Brown, 1989).  However, this has
 not been proven.  Through degradation of the solubilized
 constituents, microbes may also increase contaminant mass
 transfer rates from the NAPL by creating steeper
 concentration gradients than solubilization alone.

 Many aqueous-phase compounds dissolved from LNAPL
 sources are amenable to degradation by naturally occurring
 microorganisms in the subsurface (Norris et al., 1994; Wilson
 et al., 1986). While  in situ biodegradation occurs naturally at
 most sites, the overall rate of the reaction may be limited by a
 lack of nutrients, electron  acceptors,  or both (Thomas and
 Ward, 1989). Therefore, in situ biodegradation projects
 attempt to reduce limitations by injecting necessary nutrients
 and electron acceptors into the contaminated zone and
 stimulating naturally-occurring microorganisms.  Effectiveness
 of this approach may be limited by the inability to deliver
 nutrients and electron acceptor into heterogeneously
 contaminated and low permeability materials.  Although
 limitations exist, biodegradation of aqueous-phase
 contaminants derived from many LNAPL sources, such as
 petroleum products, is a process that is potentially applicable
 as one component of site  management at many LNAPL sites.
 The greatest utility of enhanced biodegradation may be as a
 polishing step following removal, to the extent practicable, of
 mobile and residual  NAPL. More comprehensive discussions
 of the application of  biodegradation to contaminant removal
 from soils and ground water are available (Norris et al., 1994;
 Sims et al., 1989;  Sims et al., 1992).

 Ground-Water Pump-and-Treat

 Traditional pump-and-treat systems extract contaminated
 ground water for above-ground treatment. Such systems
 have primarily been  designed to recover aqueous-phase
 contaminants. Flushing of hundreds or thousands of pore
 volumes of ground water may be required to significantly
 diminish contaminant levels at some sites (e.g., Borden and
 Kao, 1992; Geller and Hunt, 1993; Hunt et al., 1988a; Newell
 et. al., 1990).  Complete mobilization of LNAPL trapped below
the water table using increased hydraulic gradients alone is
 not practical under conditions encountered in  the field (Hunt et
al., 1988a; Wilson and Conrad, 1984). Many of the more
soluble LNAPL components may continue to dissolve in
ground water resulting in ground-water contamination at
unacceptable concentrations, potentially necessitating
 containment operations.  Depending on site conditions, time
 frames of many decades or centuries may be required to
 remove LNAPLs trapped in the saturated zone using
 dissolution alone.

 Despite the limitations, ground-water extraction and injection
 technology will be an applicable component of the overall
 remediation strategy at most sites.  Pump-and-treat systems
 may be most useful for establishing hydrodynamic control to
 prevent contaminant migration and for remediation  of
 aqueous-phase contamination in some situations in which
 contaminant sources,  including NAPLs, have been  removed
 or isolated.  However, many factors, including subsurface
 heterogeneities, may limit the effectiveness of contaminant
 removal.

 Physical Barriers

 Low permeability barriers (e.g., grout curtains, slurry walls,
 sheet piling) for control of ground-water and LNAPL flow
 (Mitchell and van Court, 1992) may be applicable as
 components of remedial operations at many sites.  Potential
 uses include containment of ground-water and/or mobile
 LNAPL during remediation. However, several of the concerns
 cited regarding difficulties in DNAPL containment (Huling and
 Weaver, 1991) will exist for LNAPL containment.  Concerns
 regarding difficulty  in assessing barrier integrity and materials
 compatibility issues should always be considered.

 Permeable treatment walls (e.g., Brown et al., 1992; Gillham
 and Burris, 1992) are an emerging technology for passive
 control of aqueous-phase contaminant migration. The
 technology is in its  infancy  with very few field-scale  trials
 reported to date. Conceptually, reactive materials or
 substances creating a reactive zone  are placed to form a
 vertical wall (Figure 15). The reactive materials remove or
 transform dissolved contaminants in ground water that passes
 through the wall. Low permeability barriers  may be
 incorporated to channel ground water through the reactive
             Low Permeability
                Barrier      ^^
         Low
      Concentration
        Plume
                             High Concentration
                            Aqueous-Phase Plume
r*\       *
>w Permeability  %
                                      Ground-Water Flow
             Low Permeability
                 Barrier
Figure 15. Map view of reactive treatment wall and low
         permeability barriers used to channel ground-water
         flow.
                                                        22

-------
wall. A biological treatment wall, for example, may slowly
allow dissolution of oxygen and nutrients into ground water,
encouraging in situ biodegradation to proceed at an
accelerated pace.

Treatment Train

Remediation may require the use of more than one
technology.  It is likely that several remediation techniques,
used in series and/or parallel applications, will be required for
maximum contaminant removal. This collaborative effort may
be referred to as  a treatment train  approach (U.S.EPA,
1992a).  A conceptual example of  a treatment train which
might be effective at an LNAPL site includes use of
conventional pumping technology  for mobile LNAPL removal.
This phase might be followed by vapor extraction for removal
of residual LNAPL and possibly coupled with ground-water
extraction to lower the water table for increased contaminant
removal. Additional technologies such as bioremediation
might be used to  further reduce contaminant concentrations.

Optimum sequencing of remedial actions in a treatment train
will be site specific and will depend on such factors as LNAPL
migration rates and distribution, remedial objectives, and
applicable remedial technologies.  Containment of migrating
LNAPL may  be an appropriate objective of  initial actions at
many sites.  Removal of LNAPL to the extent practicable will
also be an objective during early stages of  remediation at
many sites.

The treatment train concept  acknowledges the strengths and
weaknesses of various remediation strategies and couples
promising technologies to overcome limitations. A successful
treatment train will require a  thorough understanding of the
hydrogeologic and geochemical characteristics of the site.
Detailed site characterization efforts and an in-depth
understanding of the processes which affect the transport and
fate of LNAPLs in the subsurface will permit the optimization
of all possible remedial actions, maximize predictability of
remediation effectiveness, minimize remediation costs, and
make cost estimates more reliable (API, 1989; U.S.EPA,
1992a; Wilson etal., 1986).
REFERENCES

Abdul, A.S., 1988. Migration of petroleum products through
sandy hydrogeologic systems, Ground Water Montt. Rev.,
8(4): 73-81.

Abdul, A.S., 1992. A new pumping strategy for petroleum
product recovery from contaminated hydrogeologic systems:
Laboratory and field evaluations, Ground Water Monit. Rev.,
12(1): 105-114.

Abdul, A.S., S.F. Kia, and T.L. Gibson, 1989. Limitations of
monitoring wells for the detection and quantification of
petroleum products in soils and  aquifers, Ground Water Monit.
Rev., 9(2): 90-99.

Anderson, M.R.,1988. The dissolution and transport of dense
nonaqueous phase liquids in saturated porous media. Ph.D.
Dissertation, Oregon Graduate Center, Beaverton, OR,
260 pp.
Anderson, W.G., 1986. Wettability literature survey -- part 1:
Rock/oil/brine interactions and the effects of core handling on
wettability. J. Petroleum Technology, October, 1125-1144.

API (American Petroleum Institute), 1988. Phase separated
hydrocarbon contaminant modeling for corrective action, Publ.
4474, API, Washington, DC, 125 pp.

API (American Petroleum Institute), 1989. A guide to the
assessment and remediation of underground  petroleum
releases, Publ. 1628, API, Washington, DC, 81 pp.

Augustijin, D.C.M., L.S.  Lee, R.E. Jessup, and P.S.C. Rao,
1992. Remediation of soils contaminated with oily wastes:
Experimental and theoretical basis for the use of cosolvents,
in Proc. Subsurface Restoration Conf., Dallas, TX,  Rice Univ.,
Dept. of Environ. Sci. &  Eng., Houston, TX, 49-50.

Ballestero, T.P., F.R. Fiedler, and N.E. Kinner, 1994. An
investigation of the relationship between actual and apparent
gasoline thickness in a uniform sand aquifer,  Ground Water,
32(5): 708-718.

Banerjee, S., 1984. Solubility of organic mixtures in water,
Environ. Sci. Technol., 18(8): 587-591.

Bear, J., 1972. Dynamics of Fluids in Porous  Media, American
Elsevier Publishing Co., New York, 763 pp.

Benson, R.C., R.A. Glaccum, and M.R.  Noel, 1982.
Geophysical Techniques for Sensing Buried Wastes and
Waste Migration, Natl. Ground Water Assoc., Dublin, OH,
236 pp.

Blake, S.B., and R.A. Hall, 1984. Monitoring petroleum spills
with wells: Some problems and solutions, in Proc. Fourth  Natl.
Symp. on Aquifer Restoration and Ground Water Monitoring,
Natl. Ground Water Assoc., Dublin, OH, 305-310.

Blake, S.B., and R.W. Lewis, 1983. Underground oil recovery,
Ground Water Monit. Rev., 3(2): 40-46.

Borden, R.C., and C.M.  Kao, 1992. Evaluation of groundwater
extraction for remediation of petroleum-contaminated aquifers,
Water Environ. Res., 64(1): 28-36.

Boyd, G.R., and K.J. Farley, 1992. NAPL removal from
groundwater by alcohol  flooding: Laboratory studies and
applications, in Hydrocarbon Contaminated Soils and
Groundwater, Volume 2, edited  by E.J. Calabrese and P.T.
Kostecki, Lewis Publishers, Boca Raton, FL, 437-460.

Brown, M.J., D.R. Burris, J.A. Cherry, and D.M. Mackay, 1992.
Enhancement of organic contaminant retardation by the
modification of aquifer material with cationic surfactants, in
Proc. Subsurface Restoration Conf., Dallas, TX, Rice Univ.,
Dept. of Environ. Sci. &  Eng., Houston, TX, 194-196.

Brown, R., 1994. Treatment of petroleum hydrocarbons in
ground water by air sparging, in Handbook of Bioremediation,
Lewis Publishers, Boca  Raton, FL, 257 pp.
                                                        23

-------
Charbeneau, R.J., N. Wanakule, C.Y. Chiang, J.P. Nevin, and
C.L. Klein, 1989. A two-layer model to simulate floating free
product recovery: Formulation and applications, in Proc. Conf.
on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water: Prevention, Detection, and Restoration, Natl.
Ground Water Assoc., Dublin, OH, 333-345.

Charbeneau, R.J., P.B. Bedient, and R.C. Loehr, eds., 1992.
Groundwater Remediation, Water Qua). Mgmt. Library, Vol. 8,
Technomic Publishing Co., Lancaster, PA, 185 pp.

Chiang, C.Y., J.P. Nevin, and R.J. Charbeneau, 1990. Optimal
free hydrocarbon recovery from a single pumping well, in
Proc. Conf. on Petroleum  Hydrocarbons and Organic
Chemicals in Ground Water: Prevention, Detection, and
Restoration, Natl. Ground  Water Assoc., Dublin, OH,  161-178.

Chiang, C.Y., K.R. Loos, and R.A. Klopp, 1992. Field
determination of geological/chemical properties of an aquifer
by cone penetrometry and headspace analysis, Ground
Water, 30(3): 428-436.

Cohen, R.M., and J.W. Mercer,  1993. DNAPL Site Evaluation,
CRC Press, Boca Raton, FL, 314 pp.

Cohen, R.M., A.P. Bryda,  ST. Shaw,  and C.P. Spalding,
1992. Evaluation of visual methods to detect NAPL in soil and
water, Ground Water Monit. Rev., 12(4):  132-141.

Davis,  E.L., and B.K. Lien, 1993. Laboratory study on the  use
of hot water to recover light oily  wastes from sands, EPA/600/
R-93/021, U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK,
59pp.

Demond, A.H., and P.V. Roberts, 1991. Effect of interfacial
forces on two-phase capillary pressure-saturation
relationships, Water Resour. Res., 27(3): 423-437.

DePaoli, D.W., and N.J. Hutzler, 1992. Field test of
enhancement of soil venting by  heating, in Proc. Symp. on
Soil Venting, EPA/600/R-92/174, U.S.EPA, R.S. Kerr Environ.
Res. Lab., Ada, OK, 173-192.

De  Pastrovich, T.L,  Y. Baradat, R. Barthel, A. Chiarelli, and
D.R. Fussell, 1979. Protection of ground water from oil
pollution,  CONCAWE, The Hague, 61 pp.

Devrtt,  D.A., R.B. Evans, W.A. Jury, and T.H. Starks,  1987.
Soil Gas Sensing For Detection  and Mapping of Volatile
Organics, Natl. Ground Water Assoc., Dublin, OH, 270 pp.

DiGiulb, D.C., 1992. Evaluation of soil venting application,
Ground Water Issue, EPA/540/S-92/004, U.S.EPA, R.S. Kerr
Environ. Res. Lab, Ada, OK, 7 pp.

DiGiulto, D.C., and J.S. Cho, 1990. Conducting field tests for
evaluation of soil vacuum extraction application, in Proc.
Fourth Natl. Outdoor Action Conf. on Aquifer Restoration,
Ground Water Monitoring, and Geophysical Methods, Natl.
Ground Water Assoc., Dublin, OH, 587-601.

Durnford, D., J. Brookman, J. Billica, and J. Milligan, 1991.
LNAPL distribution in a conesionless soil: A field investigation
and cryogenic sampler, Ground  Water Monit. Rev., 11(3): 115-
122.
Evans, O.D., and G.M. Thompson, 1986. Field and
interpretation techniques for delineating subsurface petroleum
hydrocarbon spills using soil gas analysis, in Proc. NWWA/API
Conf. on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water: Prevention, Detection, and Restoration, Natl.
Ground Water Assoc., Dublin, OH, 444-455.

Farr, A.M., R.J. Houghtalen, and D.B. McWhorter, 1990.
Volume estimation of light nonaqueous phase liquids in
porous media, Ground Water, 28(1): 48-56.

Faust, C.R., J.H. Guswa, and J.W. Mercer, 1989. Simulation
of three-dimensional flow of immiscible fluids within and below
the unsaturated zone, Water Resour. Res., 25(12): 2449-
2464.

Feenstra,  S., D.M. Mackay, and J.A. Cherry, 1991. A method
for assessing residual NAPL based on organic chemical
concentrations in soil samples, Ground Water Monit. Rev.,
11 (2): 128-136.

Ferrand, L.A., P.C.D. Milly, and G.F. Pinder, 1989.
Experimental determination of three-fluid saturation profiles in
porous media, J. Contam. Hydrol., 4: 373-395.

Fountain, J.C., 1992. Removal of non-aqueous phase liquids
using surfactants, in Proc. Subsurface Restoration Conf.,
Dallas, TX, Rice Univ., Dept. of Environ. Sci. & Eng., Houston,
TX, 36-37.

Fulton, D.E., G.J. Reuter, and T.E. Buscheck, 1991. Hot water
enhanced recovery of phase separated lubricating oil, in Proc.
Conf. on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water: Prevention, Detection, and Restoration, Natl.
Ground Water Assoc., Dublin, OH, 143-156.

Geller, J.T., and J.R. Hunt, 1993. Mass transfer from
nonaqueous phase organic liquids in water-saturated porous
media, Water Resour. Res., 29(4): 833-845.

Gillham, R.W., and  D.R. Burris, 1992. In situ treatment walls -
Chemical dehalogenation, denitrification, and bioaugmenta-
tion, in  Proc. Subsurface Restoration Conf., Dallas, TX,  Rice
Univ., Dept. of Environ. Sci. & Eng., Houston, TX, 66-68.

Gruszczenski, T.S., 1987. Determination of a realistic estimate
of the actual formation product thickness using monitor wells:
A field bailout test, in Proc. Conf. on Petroleum Hydrocarbons
and Organic Chemicals  in Ground Water: Prevention,
Detection, and Restoration, Natl. Ground Water Assoc.,
Dublin, OH, 235-253.

Hall, R.A., S.B. Blake, and S.C. Champlin, Jr., 1984.
Determination of hydrocarbon thicknesses in sediments using
borehole data, in Proc. Fourth Natl. Symp. on Aquifer
Restoration and Ground Water Monitoring, Natl. Ground
Water Assoc., Dublin, OH, 300-304.

Hampton,  D.R., and H.G. Heuvelhorst, 1990. Designing gravel
packs to improve separate-phase hydrocarbon recovery:
Laboratory experiments, in Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and Restoration, Natl. Ground Water
Assoc, Dublin, OH,  195-209.
                                                        24

-------
Hampton, D.R., and P.D.G. Miller, 1988. Laboratory
investigation of the relationship between actual and apparent
product thickness in sands, in Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and Restoration, Natl. Ground Water
Assoc., Dublin, OH, 157-181.

Hampton, D.R., R.B. Wagner, and H.G. Huevelhorst, 1990. A
new tool to measure petroleum thickness in shallow aquifers,
in Proc. Fourth Natl. Outdoor Action Conf. on Aquifer
Restoration, Ground Water Monitoring and Geophysical
Methods, Natl. Ground Water Assoc., Dublin, OH.

Hampton, D.R., M.M. Smith, and S.J. Shank, 1991, Further
laboratory studies of gravel pack design for hydrocarbon
recovery wells, in Proc. Conf. on Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection,
and Restoration, Natl. Ground Water Assoc, Dublin, OH, 615-
629.

Hayes, D., E.G. Henry, and S.M. Testa, 1989. A practical
approach to shallow petroleum hydrocarbon recovery, Ground
Water Monit. Rev., 9(1): 180-185.

Holzer, T.L, 1976. Application of ground-water flow theory to
a subsurface oil spill, Ground Water, 14(3): 138-145.

Hudak, P.F., K.M. Clements, and H.A. Loaiciga, 1993. Water-
table correction factors applied to gasoline contamination, J.
Environ. Eng., 119(3): 578-584.

Hughes, J.P.,  C.R. Sullivan, and R.E. Zinner, 1988. Two
techniques for determining the true hydrocarbon thickness in
an unconfined sandy aquifer, in Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and  Restoration, Natl. Ground Water
Assoc., Dublin, OH, 291-314.

Huling, S.G., and J.W. Weaver, 1991. Dense nonaqueous
phase liquids, Ground Water Issue, EPA/540/4-91-002,
U.S.EPA, R.S. Kerr Environ. Res.  Lab., Ada, OK, 21 pp.

Hunt, J.R., N.  Sitar, and K.S. Udell, 1988a. Nonaqueous
phase liquid transport and cleanup, 1: Analysis of
mechanisms, Water Resour. Res., 24(8): 1247-1258.

Hunt. J.R., N.  Sitar, and K.S. Udell, 1988b. Nonaqueous
phase liquid transport and cleanup, 2: Experimental studies,
Water Resour. Res., 24(8): 1259-1269.

Hunt, W.T., J.W. Wiegand, and J.D. Trompeter, 1989. Free
gasoline thickness in monitoring wells related to ground water
elevation change, in Proc. Conf. on New Field Techniques for
Quantifying the Physical and Chemical Properties of Hetero-
geneous Aquifers, Natl. Ground Water Assoc., Dublin, OH,
671-692.

Huyakorn, P.S., Y.S. Wu, and S.Panday, 1992. A
comprehensive three-dimensional numerical model for
predicting the fate of petroleum  hydrocarbons in the
subsurface, in Proc. Conf. on Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection,
and Restoration, Natl Ground Water Assoc., Dublin, OH, 239-
253.
Ji, W., A. Dahmani, D.P. Ahlfeld, J.D. Lin, and E. Hill III, 1993.
Laboratory study of air sparging: Air flow visualization,
Ground Water Monrt. and Rem., 13(4): 115-126.

Johnson, Jr., L.A., and A.P. Leuschner, 1992. The CROW
process and bbremediation for in situ treatment of hazardous
waste sites, in Hydrocarbon Contaminated Soils and
Groundwater, Volume 2, edited by E.J. Calabrese and P.T.
Kostecki, Lewis Publishers, Boca Raton, FL, 343-357.

Johnson, P.C., M.W. Kemblowski, and J.D. Colthart, 1990a.
Quantitative analysis for the cleanup of hydrocarbon-
contaminated soils by in-situ soil venting, Ground Water,
28(3): 413-429.

Johnson, P.C., C.C. Stanley, M.W. Kembbwski, D.L. Byers,
and J.D. Colthart,  1990b. A practical approach to the design,
operation, and monitoring of in situ soil-venting systems,
Ground Water Monrt. Rev., 10(2): 159-178.

Johnson, R.L., W. Bagby, M. Perrott, and C. Chen, 1992.
Experimental  examination of integrated soil vapor extraction
techniques, in Proc. Conf. on Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection,
and Restoration, Natl. Ground Water Assoc., Dublin, OH, 441-
452.

Johnson, R.L., P.C. Johnson, D.B. McWhorter, R.E. Hinchee,
and I. Goodman, 1993. An overview of in situ air sparging,
Ground Water Monrt. and Rem., 13(4): 127-135.

Kaluarachchi, J.J., and J.C. Parker, 1989. An efficient finite
element model for modeling  multiphase flow in porous media,
Water Resour. Res., 25(1): 43-54.

Kaluarachchi, J.J., J.C. Parker, and R.J. Lenhard, 1990. A
numerical model for areal migration of water and light
hydrocarbon in unconfined aquifers, Adv. in Water Resour.,
13:29-40.

Kampbell, D.H., 1992. Subsurface remediation at a gasoline
spill site using a biovent approach, in Proc. Symp. on Soil
Venting, EPA/600/R-92/174, U.S.EPA, R.S.  Kerr Environ.
Res. Lab., Ada, OK, 309-315.

Katyal, A.K., J.J. Kaluarachchi, and J.C. Parker, 1991.
MOFAT: A two-dimensional finite element program for
multiphase  flow and multicomponent transport, program
documentation and user's guide, EPA-600/2-91-020, NTIS
PB91-191692, U.S.EPA, R.S. Kerr Environ.  Res. Lab., Ada,
OK, 109pp.

Keech, D.A., 1988. Hydrocarbon thickness on groundwater by
dielectric well logging, in  Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and Restoration, Natl. Ground Water
Assoc., Dublin, OH, 275-290.

Kemblowski, M.W., and C.Y. Chiang, 1990. Hydrocarbon
thickness fluctuations in monitoring wells, Ground Water,
28(2): 244-252.
                                                        25

-------
 Kerfoot, H.B., 1988. Is soil-gas analysis an effective means of
 tracking contaminant plumes in ground water? What are the
 limitations of the technology currently employed?, Ground
 Water Monit. Rev., 8(2): 54-57.

 Keys, W.S., 1989. Borehole Geophysics Applied to Ground
 Water Investigations, Natl. Ground Water Assoc., Dublin, OH,
 313pp.

 Keys, W.S. and L.M. MacCary, 1971.  Application of borehole
 geophysics to water-resource investigations, in Techniques of
 Water-Resource Investigations of the United States
 Geological Survey, Book 2, U.S. Govt. Printing Off.,
 Washington, D.C., 126 pp.

 King, T.V.V., and G.R. Olhoeft, 1989. Mapping organic
 contamination by detection of clay-organic processes, in Proc.
 Conf. on Petroleum Hydrocarbons and Organic Chemicals in
 Ground Water: Prevention, Detection,  and Restoration, Natl.
 Ground Water Assoc., Dublin, OH, 627-640.

 Leinonen, P.J., and D. Mackay, 1973.  The multicomponent
 solubility of hydrocarbons in water, Can. J. Chem. Eng., 51:
 230-233.

 Lenhard, R.J., and J.C. Parker, 1990. Estimation of free
 hydrocarbon volume from fluid levels in monitoring wells,
 Ground Water, 28(1): 57-67.

 Loden, M.E., 1992. Project summary: A technology
 assessment of soil vapor extraction and air sparging, EPA/
 600/SR-92/173, U.S.EPA, Risk Red. Eng. Lab., Cincinnati,
 OH, 3 pp.

 Looney, B.B., D.S. Kaback, and J.C. Corey, 1992.
 Environmental  restoration using horizontal wells: A field
 demonstration, in Proc. Subsurface Restoration Conf., Dallas,
 TX, Rice Univ., Dept. of Environ. Sci. & Eng., Houston, TX,
 41-43.

 Lyman, W.J., and D.C. Noonan, 1990. Assessing UST
 corrective action technologies: Site assessment and selection
 of unsaturated zone treatment technologies, EPA/600/2-90/
 011, U.S.EPA,  Risk Red. Eng. Lab., Cincinnati, OH, 107pp.

 Marley, M.C., D.J. Hazebrouck, and M.T. Walsh,  1992a. The
 application of in situ air sparging as an innovative soils and
 ground water remediation technology, Ground Water Monit.
 Rev., 12(2): 137-145.

 Marley, M., D. Hazebrouck, and M. Walsh, 1992b. Tiny
 bubbles pop to deep clean, Soils, October: 8-12.

 Marrin, D.L., 1988. Soil-gas sampling and misinterpretation.
Ground Water Monrt. Rev., 8(2): 51 -54.

 Marrin, D.L, and H.B. Kerfoot, 1988. Soil-gas surveying
techniques, Environ. Sci. Technol., 22(7): 740-745.

 Mendoza, C.A., and T.A. McAlary, 1989.  Modeling of
groundwater contamination caused by organic solvent vapors,
Ground Water, 28(2): 199-206.
 Mercer, J.W., and R.M. Cohen, 1990. A review of immiscible
 fluids in the subsurface: Properties, models, characterization,
 and remediation, J. Contam. Hydrol., 6:107-163.

 Miller, C.T., M.M. Poirier-McNeill, and A.S. Mayer, 1990.
 Dissolution of trapped nonaqueous phase liquids: Mass
 transfer characteristics, Water Resour. Res., 26(11): 2783-
 2796.

 Mitchell, U.K., and W.A. van Court, 1992. Barriers - walls and
 covers,  in Proc. Subsurface Restoration Conf., Dallas, TX,
 Rice Univ., Dept. of Environ. Sci. & Eng., Houston, TX, 34-35.

 Morgan, J.H., 1992. Horizontal drilling applications of
 petroleum technologies for environmental purposes, Ground
 Water Monit. Rev., 12(3): 98-102.

 Newell,  C.J., J.A. Conner, and O.K. Wilson, 1990. Pitot test for
 evaluating the effectiveness of enhanced in-situ
 biodegradation for soil remediation, in Proc. Conf. on
 Petroleum Hydrocarbons and Organic Chemicals in Ground
 Water: Prevention, Detection, and Restoration, Natl. Ground
 Water Assoc., Dublin, OH, 369-383.

 Norris, R.D., R.E. Hinchee, R. Brown, P.L. McCarty, L.
 Semprini, J.T. Wilson, D.H. Kampbell, M. Reinhard, E.J.
 Bouwer, R.C. Borden, T.M. Vogel, J.M.  Thomas, and C.H.
 Ward, 1994. Handbook of Bioremediation,  Lewis Publishers,
 Boca Raton, FL, 257  pp.

 Oakley, D., M. Thacker, K. Singer, J.  Koelsch, and B. Mabson,
 1992. The use of horizontal wells in remediating and
 containing a jet fuel plume-preliminary findings, in Proc. Conf.
 on Petroleum Hydrocarbons and Organic Chemicals in
 Ground  Water: Prevention, Detection, and Restoration, Natl.
 Ground  Water Assoc., Dublin, OH, 1-13.

 Olhoeft, G.R., 1986. Direct detection of  hydrocarbon and
 organic chemicals with ground penetrating radar and complex
 resistivity, in Proc. NWWA/API Conf. on Petroleum
 Hydrocarbons and Organic Chemicals in Ground Water:
 Prevention, Detection, and Restoration, Natl. Ground Water
 Assoc.,  Dublin, OH, 284-305.

 Ostendorf, D.W., E.E. Moyer, R.J. Richards, E.S. Hinlein,
 Y. Xie, and R.V. Rajan, 1992. LNAPL distribution and
 hydrocarbon vapor transport in the capillary fringe, EPA/600/
 R-92/247, U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK,
 122 pp.

 Pantazidou, M.,  and N. Sitar, 1993. Emplacement of
 nonaqueous liquids in the vadose zone, Water Resour. Res
29(3): 705-722.

 Pedersen, T.A.,  and J.T. Curtis, 1991. Soil vapor extraction
technology, Reference Handbook, EPA/540/2-91/003,
 U.S.EPA, Risk Red. Eng. Lab., Cincinnati, OH, 316 pp.

Piwoni, M.D., and J.W. Keeley, 1990. Basic concepts of
contaminant sorption  at hazardous waste sites, Ground Water
 Issue, EPA/540/4-90/053, U.S.EPA, R.S. Kerr Environ.  Res.
Lab., Ada, OK, 7 pp.
                                                        26

-------
Powers, S.E., C.O. Loureiro, LM. Abriola, and W.J. Weber,
Jr., 1991. Theoretical study of the significance of
nonequilibrium dissolution of nonaqueous phase liquids in
subsurface systems, Water Resour. Res., 27(4): 463-477.

Powers, S.E., LM. Abriola,  and W.J. Weber, Jr., 1992. An
experimental investigation of nonaqueous phase liquid
dissolution in saturated subsurface systems: Steady state
mass transfer rates, Water Resour. Res., 28(10): 2691-2705.

Rao, P.S.C., LS. Lee, and A.L Wood, 1991. Solubility,
sorption, and transport of hydrophobic organic chemicals in
complex mixtures, Environmental Research Brief, EPA/600/M-
91/009, U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK,
14pp.

Reidy,  P.J., W.J. Lyman, and D.C. Noonan, 1990. Assessing
UST corrective action technologies: Early screening of
cleanup technologies for the saturated zone, EPA/600/2-90/
027, U.S.EPA, Risk Red. Eng. Lab., Cincinnati, OH, 124 pp.

Schwille, F., 1988. Dense Chlorinated Solvents in Porous and
Fractured Media, Lewis  Publishers, Chelsea, Ml, 146 pp.

Seitz, W.R., 1990. In situ detection of contaminant plumes in
ground water, Special Report 90-27,  U.S. Army Corps of
Engineers, Cold Regions Research & Engineering Laboratory,
12pp.

Sims, J.L., R.C. Sims, and J.E. Matthews, 1989.
Bioremediation of contaminated surface soils, EPA 600/9-89/
073, U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK, 23 pp.

Sims, J.L., J.M.  Suflita, and H.H. Russell, 1992. In-situ
bioremediation of contaminated ground water, Ground Water
Issue, EPA/540/S-92/003, U.S.EPA,  R.S. Kerr Environ. Res.
Lab., Ada, OK, 11 pp.

Sims, R.C., 1990. Soil remediation techniques at uncontrolled
hazardous waste sites, J. Air Waste Manage. Assoc., 40(5):
704-730.

Srttler,  S.P., G.L. Swinford, and D.G. Gardner, 1992. Use of
thermal-enhanced soil vapor extraction to accelerate
remediation of diesel-affected soils, in Proc. Conf. on
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water: Prevention, Detection, and Restoration, Natl. Ground
Water Assoc., Dublin, OH, 413-426.

Smolley,  M., and J.C. Kappmeyer, 1991. Cone penetrometer
tests and HydroPunch sampling: A screening technique for
plume definition, Ground Water Monit. Rev.,  11 (2):  101 -106.

Sullivan, C.R., R.E. Zinner, and J.P. Hughes, 1988. The
occurrence of hydrocarbon on an unconfined aquifer and
implications for liquid recovery, in Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water
Prevention, Detection, and  Restoration, Natl. Ground Water
Assoc., Dublin, OH, 135-156.

Testa,  S.M., and M.T. Paczkowski, 1989. Volume
determination and recoverability of free hydrocarbon, Ground
Water Monit. Rev., 9(1): 120-128.
Testa, S.M., D.L. Winegardner, and C.B. Burris, 1992.
Reinjection of coproduced groundwater in relation to LNAPL
occurrence, in Proc. Conf. on Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection,
and Restoration, Natl. Ground Water Assoc., Dublin, OH, 127-
145.

Thomas, T.M., and C.H. Ward, 1989. In situ biorestoratbn of
organic contaminants in the subsurface, Environ. Sci.
Technol., 23: 760-765.

Thompson, G.M., and D.L. Marrin, 1987. Soil gas contaminant
investigations: A dynamic approach, Ground Water Monit.
Rev., 7(3): 88-93.

Travis, C.C., and J.M. Macinnis, 1992.  Vapor extraction of
organics from subsurface soils-Is it effective?, Environ. Sci.
Technol., 26(10): 1885-1887.

Udell, K.S., 1992. Thermally enhanced removal of the oily
phase, in Proc. Subsurface Restoration Conf., Dallas, TX,
Rice Univ., Dept. of Environ. Sci. & Eng., Houston, TX, 51-53.

U.S.EPA, 1988. Cleanup of releases from petroleum USTs:
Selected technologies, EPA/530/UST-88/001, U.S.EPA,
Washington, DC, 110pp.

U.S.EPA, 1990. Subsurface contamination reference guide,
EPA/540/2-90/011, U.S.EPA, Washington, DC,  13 pp.

U.S.EPA, 1991. Seminar publication: Site characterization for
subsurface remediation, EPA/625/4-91/026, U.S.EPA, R.S.
Kerr Environ.  Res.  Lab., Ada, OK, 259  pp.

U.S.EPA, 1992a. Dense  nonaqueous phase liquids -- A
workshop summary, Dallas, Texas, April 16-18, 1991, EPA/
600/R-92/030, U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada,
OK, 81 pp.

U.S.EPA, 1992b. Estimating potential for occurrence of
DNAPL  at Superfund sites, Publ. 9355.4-07FS,  U.S.EPA, R.S.
Kerr Environ.  Res.  Lab., Ada, OK, 9 pp.

U.S.EPA, 1992c. Proceedings of the symposium on soil
venting, April  29 - May 1, 1991, Houston, Texas, EPA/600/R-
92/174,  U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK, 334
PP-

U.S.EPA, 1993a. Accutech pneumatic  fracturing extraction
and hot  gas injection, Phase 1, EPA/540/AR-93/509,
U.S.EPA, Risk Red. Eng. Lab., Cincinnati, OH, 44 pp.

U.S.EPA, 1993b. Hydraulic fracturing technology, Applications
analysis and technology evaluation report, EPA/540/R-93/505,
U.S.EPA, Risk Red. Eng. Lab., Cincinnati, OH, 119 pp.

U.S.EPA, 1994. Manual: Alternative methods for fluid delivery
and recovery, EPA/625/R-94/003, U.S.EPA, Risk Red. Eng.
Lab., Cincinnati, OH, 87 pp.

van Dam, J., 1967. The migration of hydrocarbons in a water
bearing  stratum, in The Joint Problems of the Oil and Water
Industries, P.  Hepple, ed., Elsevier, Amsterdam, 55-96.
                                                        27

-------
van der Heijde, P.K.M., and O.A. Elnawawy, 1993.
Compilation of ground-water models, EPA/600/2-93/118,
U.S.EPA, R.S. Kerr Environ. Res. Lab., Ada, OK, 279 pp.

van Eyk, J., 1992. Bioventing: An in situ remedial technology,
scope and limitations, in Proc. Symp. on Soil Venting, EPA/
600/R-92/174, U.S. EPA, R.S.Kerr Environ. Res. Lab., Ada,
OK, 317-334.

Wagner, R.B., D.R. Hampton, and J.A. Howell, 1989. A new
tool to determine the actual thickness of free product in a
shallow aquifer,  in Proc. Conf. on Petroleum Hydrocarbons
and Organic Chemicals in Ground Water: Prevention,
Detection, and Restoration, Natl. Ground Water Assoc.,
Dublin, OH, 45-59.

Walther, E.G., A.M. Pitchford, and G.R. Olhoeft, 1986. A
strategy for detecting subsurface organic contaminants, in
Proc. NWWA/API Conf. on Petroleum Hydrocarbons and
Organic Chemicals in Ground Water: Prevention, Detection,
and Restoration, Natl. Ground Water Assoc., Dublin, OH, 357-
381.

Waterloo Centre For Ground Water Research, 1989.
University of Waterloo Short Course, "Dense Immiscible
Phase  Liquid Contaminants in Porous and Fractured Media",
Krtchner,  Ontario, Canada, Nov. 6-9, 1989.

Weaver, J.W., R.J. Charbeneau, J.D. Tauxe, B.K. Lien, and
J.B. Provost,  1994,  The Hydrocarbon Spill Screening Model
(HSSM) Volume 1:  User's Guide, EPA/600/R-94/039a, U.S.
EPA, R.S. Kerr Environ. Res. Lab., Ada, OK, 212 pp.

Wickramanayake, G.B., N. Gupta, R.E. Hinchee, and B.J.
Nielsen, 1991. Free petroleum hydrocarbon volume estimates
from monitoring  well data, J. Environ. Eng., 117(5): 686-691.

Williams, D.E., and D.G. Wilder, 1971.  Gasoline pollution of  a
ground-water reservoir - A case history, Ground Water, 9(6):
50-54.

Wilson, J.L., and S.H. Conrad, 1984. Is physical displacement
of residual hydrocarbons a realistic possibility in aquifer
restoration?, in Proc. NWWA/API Conf. on Petroleum
Hydrocarbons and Organic Chemical in Ground Water:
Prevention, Detection, and Restoration, Natl. Ground Water
Assoc., Dublin, OH, 274-298.

Wilson, J.T., L.E. Leach,  M. Henson and J.N. Jones, 1986. In
situ biorestoratbn as a ground water remediation technique,
Ground Water Monrt. Rev., 6(4): 56-64.

Wilson, S.B., and R. A. Brown, 1989. In situ bioreclamation:  A
cost effective technology to remediate subsurface organic
contamination, Ground Water Monrt. Rev., 9(1): 175-179.

Yaniga, P.M., 1984. Hydrocarbon retrieval and apparent
hydrocarbon thickness: Interrelationships to recharging/
discharging aquifer conditions, in Proc. NWWA/API Conf. on
Petroleum Hydrocarbons and Organic Chemicals in Ground
Water: Prevention, Detection and Restoration, Natl. Ground
Water Assoc., Dublin, OH, 299-329.
Yaniga, P.M., and J.G. Warburton, 1984. Discrimination
between real and apparent accumulation of immiscible
hydrocarbons on the water table: A theoretical and empirical
analysis, in Proc. Fourth Natl. Symp. and Expo, on Aquifer
Restoration and Ground Water Monitoring, Natl. Ground
Water Assoc, Dublin, OH, 311-315.

Zinner, R.E., E.A. Hodder, W.E. Carroll, and C.A. Peck, 1991.
Utilizing groundwater reinjection in the design of a liquid
hydrocarbon recovery system, in Proc. Conf. on Petroleum
Hydrocarbons and Organic Chemicals in Ground Water:
Prevention, Detection, and Restoration, Natl. Ground Water
Assoc., Dublin, OH, 469-483.
                                                        28

-------