United StatesEnvironmental ProtectionAgency

EPA/625/R-95/005July 1996

Office of Research andDevelopmentWashington DC 20460

Pump-and-TreatGround-Water Remediation

A Guide for Decision Makersand Practitioners

1

1. Introduction to Pump-and-Treat Remediation

• What is involved in “smart” application ofthe pump-and-treat approach?

• What are tailing and rebound, and how canthey be anticipated?

• What are the recommended methods formeeting the challenges of effective hydrauliccontainment?

• How can the design and operation of a pump-and-treat system be optimized and its perfor-mance measured?

• When should variations and alternatives toconventional pump-and-treat methods beused?

By presenting the basic concepts of pump-and-treat technology, this guide provides decision-makers with a foundation for evaluating theappropriateness of conventional or innovativeapproaches. An in-depth understanding ofhydrogeology and ground-water engineering isrequired, however, to design and operate a pump-and-treat system for ground-water remediation.Readers seeking more information on specifictopics covered in this booklet should refer to theU.S. Environmental Protection Agency (EPA)documents listed at the end of this guide (Section9).

Pump-and-treat is one of the most widely usedground-water remediation technologies. Conven-tional pump-and-treat methods involve pumpingcontaminated water to the surface for treatment.This guide, however, uses the term pump and treatin a broad sense to include any system wherewithdrawal from or injection into ground water ispart of a remediation strategy. Variations andenhancements of conventional pump and treatinclude hydraulic fracturing as well as chemicaland biological enhancements. The pump-and-treatremediation approach is used at about three-quarters of the Superfund sites where groundwater is contaminated and at most sites wherecleanup is required by the Resource Conservationand Recovery Act (RCRA) and state laws [Na-tional Research Council (NRC), 1994]. Althoughthe effectiveness of pump-and-treat systems hasbeen called into question (Sidebar 1), after twodecades of use, this approach remains a necessarycomponent of most ground-water remediationefforts and is appropriate for both restoration andplume containment.

This guide provides an introduction to pump-and-treat ground-water remediation by addressingthe following questions:

• When is pump and treat an appropriateremediation approach?

2

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

Pump-and-treat systems for remediating ground watercame into wide use in the early to mid-1980s. By the

early 1990s, evaluations by EPA (Keely, 1989; U.S.EPA, 1989; Haley et al., 1991) and others (Freeze andCherry, 1989; Mackay and Cherry, 1989) called into thequestion the performance of pump-and-treat systems.The general "failure" of the pump-and-treat approachwas identified as its inability to achieve "restoration"

(i.e., reduction of contaminants to levels required byhealth-based standards) in 5 to 10 years, as anticipatedin the design phase of projects. Although a variety offactors contributed to this shortcoming, tailing andrebound (Section 4) represented the major barrier toachieving remediation goals. Pump-and-treat systemswere criticized more pointedly by Travis and Doty

(1990), who asserted as a "simple fact" that "contami-nated aquifers cannot be restored through pumping andtreating."

Expectations for the effectiveness of pump-and-treattechnology, however, may have been too high. Ground-

water scientists and engineers generally agree thatcomplete aquifer restoration is an unrealistic goal formany, if not most, contaminated sites. Nonetheless,further experience with pump-and-treat systems

indicates that full restoration at some sites withrelatively simple characteristics is possible; moreover, at

many sites, full restoration of ground-water quality canbe achieved for part of a site (NRC, 1994). Forexample, Bartow and Davenport (1995), in a review of37 applications of pump-and-treat systems in SantaClara Valley, California, found that one site hadachieved maximum contaminant levels (MCLs) for all

contaminants and about one-third achieved, or werenear, MCLs for one or more parameters. Bartow andDavenport's conclusion that pump-and-treat systems hadsignificantly reduced the mass of volatile organiccontaminants (VOCs) in the region's ground waterindicates how expectations regarding the technologyhave changed.

Combining the pump-and-treat approach with in situ

bioremediation (see Section 7.4) provides furtheropportunities for improving the effectiveness of ground-water cleanup. For example, Marquis (1995) suggestedthat in situ bioremediation used with the pump-and-treat

approach should always be considered as an option forremediation of sand and gravel aquifers contaminatedwith biodegradable organic compounds, especiallyvolatile aromatic and polyaromatic hydrocarbons.

3

2. Appropriate Use of Pump-and-Treat Technology

ciently that the aquifer complies with cleanupstandards or the treated water withdrawn fromthe aquifer can be put to beneficial use.

Although hydraulic containment and cleanupcan represent separate goals, more typically,remediation efforts are undertaken to achieve acombination of both. For example, if restoration isnot feasible, the primary objective might becontainment. In contrast, where a contaminatedwell is used for drinking water but the contami-nant source has not been identified, treatment atthe wellhead might allow continued use of thewater even though the aquifer remains contami-nated.

Pump-and-treat systems are used primarily toaccomplish the following:

• Hydraulic containment. To control themovement of contaminated ground water,preventing the continued expansion of thecontaminated zone. Figure 1 illustrates threemajor configurations for accomplishinghydraulic containment: (1) a pumping wellalone, (2) a subsurface drain combined with apump well, and (3) a well within a barrierwall system.

• Treatment. To reduce the dissolved contami-nant concentrations in ground water suffi-

4

Figure 1.Examples of hydrauliccontainment in a planview and cross sectionusing pump-and-treattechnology: (a) pump well,(b) drain, and (c) wellwithin a barrier wallsystem (after Cohen etal., 1994)

991

Drain

Static Water Table

Plume

CaptureZone Limit

Static Water Table

Plume

Drain

CaptureZone Limit

(a)

(b)

Plume

998 UpgradientBarrier Wall

Plume

Downgradient Barrier Wall

CaptureZone Limit

(c)

993

992

991

998

991

992

993

994

995

996

997

Plume

99

8

99

7

99

6

99

5

99

4

998

991

992

993

994

995

996

997

Barrier Wall

Plume

994

999

99

3

99

2

99

1

5

3. Smart Pump-and-Treat Techniquesmaterial to the ground water. Vapors also mightmigrate to the water table and contaminate groundwater without infiltration.

Source removal is the most effective way toprevent further contamination. Where inorganic ororganic contaminants are confined to the vadosezone, removal is usually the preferred option.When removal is not feasible, as is often the casewith dense non-aqueous phase liquids (DNAPLs)residing below the water table, containment is anessential initial step in remediation. In somesituations, containment can be achieved throughcapping, which prevents or reduces infiltration ofrainfall through the contaminated soil. Cappingcan be ineffective if water table fluctuations occurwithin the zone of contamination or when NAPLvapors are present.

3.2. Thorough Site CharacterizationComprehensive characterization of the contami-

nated site serves two major functions:

• Accurately assessing the types, extent, andforms of contamination in the subsurfaceincreases the likelihood of achieving treat-ment goals. This requires an understanding ofthe physical phases in which contaminantsexist (mainly sorbed and aqueous phases forinorganic contaminants, and sorbed, NAPL,aqueous, and gaseous phases for organicliquids) and quantification of the distributionbetween the phases. Indeed, inadequate sitecharacterization has undermined some pump-

A fundamental component of any ground-waterremediation effort using the pump-and-treatapproach is contaminant removal or control. Thus,effective remediation of ground water usingpump-and-treat technology requires knowledge ofcontaminants and site characteristics. Addition-ally, the remediation plan should call for imple-mentation of dynamic system management basedon a statement of realistic objectives (Hoffman,1993).

3.1. Contaminant Removal/ControlAny ground-water cleanup effort will be

undermined unless inorganic and organic con-taminant sources are identified, located, andeliminated, or at least controlled, to preventfurther contamination of the aquifer. Toxicinorganic substances may serve as a continuingsource of contamination through mechanismssuch as dissolution and desorption. At manycontaminated sites, organic liquids are a majorcontributor to ground-water contamination.Figure 2 illustrates four common types of con-taminant plumes, each characterized by theliquid’s density relative to water and the degree towhich the liquid mixes with water. Even when theorganic liquid resides exclusively in the vadosezone (i.e., the area between the ground surfaceand the water table) it can serve as a source ofground-water contamination. In such situations,contamination occurs when percolating watercomes in contact with the liquid (sometimescalled product) or its vapors and carries dissolved

6

Figure 2.Contaminant plumesas a function ofdensity andmiscibility withground water:(a) light liquids(gasoline andmethanol) createcontaminant plumesthat tend to flow inthe upper portions ofan aquifer; (b) denseliquids (perchloro-ethylene [PCE] andethylene glycol)create a plume thatcontaminates the fullthickness of anaquifer (adaptedfrom Gorelick et al.,1993).

(a)

(b)

7

and-treat efforts; for instance, when after afew years greater quantities of contaminantshad been removed than were identified in theinitial site assessment.

• A thorough, three-dimensional characteriza-tion of subsurface soils and hydrogeology,including particle-size distribution, sorptioncharacteristics, and hydraulic conductivity,provides a firm basis for appropriate place-ment of pump-and-treat wells. Such informa-tion is also required for evaluating the extentto which tailing and rebound may presentproblems at a site (Section 4).

Three-dimensional characterization techniquesinclude primarily indirect observations, usingsurface and borehole geophysical instruments andcone-penetration measurements, and directsampling of soil and ground water. Importantadvances in soil sampling technology have beenmade relatively recently, such as continuoussamplers used with a hollow-stem auger (Figure3) and smaller continuous-core, direct-pushequipment that also can be used to collect ground-water samples without installing wells (Figure 4).Vibratory drilling methods are another innovativetechnique for collecting soil cores and ground-water samples. Additionally, sensitive boreholeflow meters that allow measurement of verticalchanges in hydraulic conductivity in a boreholerepresent an important recent development. Thesetechniques allow subsurface mapping to begenerated with a level of detail that generallywould be prohibitively expensive using conven-tional drilling and sampling methods. Figure 5presents a conceptual diagram of trichloroethene(TCE) contamination at a complex site(Sidebar 2) developed from extensive use ofdirect-push sampling techniques.

Moreover, if sufficient data are obtained, theinterpretation of subsurface data can be enhanced

greatly by performing two- and three-dimensionalcomputer modeling of the subsurface. Figure 6shows a conceptual model of a site developed bycombining contour visualization of a contaminantplume of benzene with subsurface lithologic logs(Sidebar 3). EPA’s SITE3D software, beingdeveloped by the National Risk ManagementLaboratory’s Subsurface Protection and Remedia-tion Division, allows three-dimensional visualiza-tion of contaminant plumes (Figure 7). Statisticalsoftware developed by EPA such as Geo-EAS andGEOPACK for geostatistical analysis and con-touring of ground-water contaminant data andGRITS/STAT for analysis of contaminant concen-trations are among the many computer-based toolsavailable for analyzing subsurface data.

3.3. Dynamic Management of theWell Extraction Field

To be effective, pump-and-treat efforts must gobeyond initial site characterization, using infor-mation gathered after remediation operations areunder way to manage the well extraction fielddynamically. For instance, information collectedwhile drilling and installing extraction wells,operating pumping wells, and tracking changes inwater levels in monitoring wells (Section 6.4) andcontaminant concentrations in observation wellscan refine the portrayal of the site.

Dynamic management of the well extractionfield based on more comprehensive informationcan provide both economic and environmentalbenefits. In general, additional information aboutthe site and the pump-and-treat effort allowsoperators to make informed decisions about theefficient use of remediation resources. Morespecifically, this flexible site management ap-proach may facilitate greater success in hydrauliccontainment (Section 5). Ultimately, the timerequired to achieve cleanup goals might beminimized.

8

Figure 3 .This hollow-stemauger is fittedwith a 5-footsampling tubethat collects acontinuous coreas the augeradvances,allowing detailedand accurateobservation ofsubsurfacelithology. Whendrilling iscompleted, amonitoring wellalso can beinstalled.

9

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

A key component of the dynamic managementapproach is the effective design and operation ofthe pump-and-treat system. The following tech-niques can be useful in this regard:

• Using capture zone analysis, optimizationmodeling, and data obtained from monitoringthe effects of initial extraction wells toidentify the best locations for wells (Section6.1).

• Phasing the construction of extraction andmonitoring wells so that information obtainedfrom operation of the initial wells informsdecisions about siting subsequent wells(Section 6.2).

• Phasing pumping rates and the operation ofindividual wells to enhance containment,avoid stagnation zones (Section 5.2.3), and

10

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

ensure removal of the most contaminatedground water first (Section 6.3).

3.4. Realistic Cleanup GoalsUnrealistic expectations for the pump-and-treat

approach can lead to disappointments in systemperformance (Sidebar 1). Indeed, a cleanup goalthat is realistic for one site may not be reasonableelsewhere. The Committee on Ground WaterCleanup Alternatives of the National Academy of

Sciences (NRC, 1994) has identified three majorclasses of sites based on hydrogeology (Sidebar2) and contaminant chemistry (see Table 1):

• Class A. Sites where full cleanup to health-based standards should be feasible usingcurrent technology. Such sites include homo-geneous single- and multiple-layer aquifersinvolving mobile, dissolved contaminants.

11

• Class B. Sites where the technical feasibilityof complete cleanup is likely to be uncertain.This class includes a wide range of hydrogeo-logic settings and contaminant types that donot fall into classes A or C.

• Class C. Sites where full cleanup of thesource areas to health-based standards is notlikely to be technically feasible. Such sitesinclude fractured-rock aquifers contaminated

by free-product light nonaqueous phaseliquids (LNAPL) or DNAPL and single- ormultiple-layered heterogeneous aquiferscontaminated by a free-product DNAPL(Sidebar 4).

Typically, preliminary ground-water cleanupefforts at contaminated sites are focused onstandards established for drinking water, such asfederal or state maximum contaminant levels

Sidebar 2Major Types of Hydrogeologic Settings

• Fractured aquifers typically consist oflow-permeability rock where most ground-waterflow is in joints and fractures.

Single-layered aquifers are less complex thanmultiple-layered aquifers, which are separated byless-permeable strata, because the possibility of cross

contamination between aquifers by either upward ordownward movement becomes a consideration. Notethat the term aquifer is used in this guide in a broadsense to include any area within the saturated zonewhere the presence of ground-water contamination is ofsufficient concern to require remediation. Figure 2

illustrates a homogenous, single-layer aquifer.Figure 15c (described in Section 4.2.5) illustrates atwo-layered homogenous aquifer. Figure 5 illustrates amultiple-layered heterogeneous aquifer. The challengeof ground-water cleanup increases along with aquifercomplexity because of difficulties in delineating

contaminant sources and pathways and the increasedlikelihood of tailing and rebound effects (Section 4).

The Committee on Ground Water Cleanup Alterna-tives of the National Academy of Sciences has definedthree major hydrogeologic settings for evaluating thetechnical feasibility of ground-water cleanup based onthe degree of uniformity of the aquifer material andlayering (Table 1).

• Homogeneous aquifers consist of materials that donot vary significantly in their water-transmittingproperties. Contaminant movement in homogeneousaquifers is largely a function of the hydraulicconductivity of the aquifer. For example, a homoge-neous aquifer might comprise permeable, well-sorted

sands or gravels.

• Heterogeneous aquifers consist of materials that varyin their water-transmitting properties laterally,vertically, or in both directions. Contaminants inheterogeneous aquifers move preferentially in thehigh-permeability zones, resulting in more rapid

transport than would be expected based on theaverage hydraulic conductivity of the aquifer. A sandaquifer with lenses of silts and clays is an example ofa heterogeneous aquifer.

12

(MCLs) or nonzero MCL goals (MCLGs). EPAhas established procedures, however, by whichefforts can target alternative goals at Superfundand RCRA sites using alternate concentrationlimits (ACLs) where ground-water discharges intonearby surface water (U.S. EPA, 1988) or demon-strating the technical impracticality (TI) ofground-water cleanup (U.S. EPA, 1993; Feldmanand Campbell, 1994). At DNAPL sites where theTI of ground-water cleanup has been demon-strated, the remedial strategy might call for

removal of as much of the DNAPL as is feasible,containment of the remaining DNAPL, andtreatment of the aqueous contaminant plumeoutside the containment area. Consequently, evenat Class C and Class B sites where restoration isnot feasible, application of some form of thepump-and-treat approach may be required eitherto help contain the contaminant source andaqueous-phase plume or to clean up the contami-nated ground water outside the containment area.

Figure 6 .GEOS computerscreen showingorganiccontaminantplume in relationto subsurfacestratigraphy (seetext fordiscussion).

13

Sidebar 3Computer Graphics as a Site Characterization Tool

Computer visualization can help focus attention onthe types of additional information needed whencharacterizing a site before initiating a remediationeffort. For example, in Figure 6 the contoured benzenedata, collected from the sand/gravel #1 aquifer (yellowin the cross section in the upper right) shows two areas

of high concentration (i.e., MW-5 and MW-7). Does thisrepresent a contaminant plume from a single source, ordoes it indicate two contaminant plumes from separatesources? An examination of the cross section (upperright of center) indicates that sand/gravel #1 aquifer atMW-7 has the lower “high” concentration of benzene,

suggesting that the two plumes might be related.

The cross section also shows, however, that theaquifer is quite thin between the two monitoring wells.Indeed, examination of monitoring well MP-10, which

lies near the cross section (see lower center log) revealsthat the aquifer is missing at this point. This suggeststhe absence of a concentration gradient between the twowells, indicating that two different sources may beinvolved.

Further analysis of the spatial distribution of the sand/

gravel #1 aquifer, including flow directions as indicatedby potentiometric heads, and possibly additionalsampling for benzene would be required to determine ifone or two sources are contributing to the contamina-tion.

This particular example also cautions against relying

exclusively on computer-generated interpolations,which can suggest features that are not actually present(i.e., continuity of the aquifer between MW-5 andMW-7).

14

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

15

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

Contaminant Chemistry

StronglyMobile Sorbed,

Dissolved Dissolved Strongly Separate Separate(degrades/ Mobile, (degrades/ Sorbed, Phase Phase

Hydrogeology volatilizes) Dissolved volatilizes) Dissolved LNAPL DNAPL

Homogeneous, A A B B B B

single layer (1) (1-2) (2) (2-3) (2-3) (3)

Homogeneous, A A B B B B

multiple layers (1) (1-2) (2) (2-3) (2-3) (3)

Heterogeneous, B B B B B C

single layer (2) (2) (3) (3) (3) (4)

Heterogeneous, B B B B B C

multiple layers (2) (2) (3) (3) (3) (4)

Fractured B B B B C C

(3) (3) (3) (3) (4) (4)

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

16

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

The light (LNAPLs) and dense (DNAPLs) immiscible

nonaqueous phase liquids shown in Figure 2 pose themost difficult problems for ground-water cleanup

because of their complex interactions with water andsolids in the subsurface. Figure 5 illustrates thesecomplexities when trichloroethene (TCE), a DNAPL hasmoved through a heterogeneous, multiple-layeredalluvial aquifer (at a site in Tennessee). Four distinctforms, or phases, of TCE are evident:

• The NAPL emits a vapor phase in the unsaturatedzone that moves by diffusion. DNAPL vapors tend tosink until they reach impermeable layers (Figure 5) orthe water table. Even if the NAPL does not reach theground water, contamination can occur by dissolutionof the vapors directly into the ground water or bywater percolating through the unsaturated zone.

• Residual NAPL remains after the free product hasmoved through the subsurface by gravity or beendisplaced by water (Figure 8a and b). Residual NAPLexists as single- to complex-shaped blobs that fill

pore spaces (Figure 8b and 9). The amount ofresidual NAPL remaining in the subsurface dependson the subsurface material and the type of NAPL.

Residual saturation in the unsaturated zone typicallyranges from 10 to 20 percent of the subsurfacevolume and in the saturated zone generally rangesfrom 10 to 50 percent (Cohen and Mercer, 1993).Figure 5 differentiates residual TCE above andbelow the water table.

• Free product exists where most of the pore space isfilled by the NAPL. It accumulates wherever abarrier prevents downward movement. LNAPLs,such as gasoline, tend to float on top of the watertable, whereas DNAPLS tend to sink until they reachan impermeable layer (Figure 5).

• Dissolved NAPL forms the aqueous contaminant

plume that moves in the direction of ground-waterflow. The residual NAPL and free product can serveas a source of ground-water contamination as long asthey remain in the subsurface.

17

Figure 8. Dark NAPL (Soltrol) and water in a homogenous micromodel after (a) the displacement of water byNAPL and then (b) the displacement of NAPL by water, with NAPL at residual saturation (Wilson et al.,1990).

a b

18

Figure 9.Photomicrographs of (a) asingle blob occupying onepore body, and (b) a doubletblob occupying two porebodies and a pore throat(Wilson et al., 1990).

a

b

19

4. Anticipating Tailing and Rebound Problems

The phenomena of tailing and rebound arecommonly observed at pump-and-treat sites.Tailing refers to the progressively slower rate ofdecline in dissolved contaminant concentrationwith continued operation of a pump-and-treatsystem (Figure 10). Rebound is the fairly rapidincrease in contaminant concentration that canoccur after pumping has been discontinued. Thisincrease may be followed by stabilization of thecontaminant concentration at a somewhat lowerlevel.

4.1. Effects of Ta iling and Reboundon Remediation Efforts

Tailing presents two main difficulties forground-water restoration:

• Longer treatment times. Without tailing,contaminants theoretically could be removedby pumping a volume of water equivalent tothe volume of the contaminant plume (Figure10). The tailing effect, however, significantlyincreases the time pump-and-treat systemsmust be operated to achieve ground-waterrestoration goals. Indeed, pumping may needto be conducted for hundreds of years ratherthan tens of years.

• Residual concentrations in excess of thecleanup standard. When tailing occurs, ofteninitially the decline in the rate of contaminantconcentrations is fairly rapid, followed by amore gradual decline that eventually stabilizesat an apparent residual concentration levelabove the cleanup standard (Figure 10).

Rebound is most problematic when a pump-and-treat system attains the cleanup standard, butconcentrations subsequently increase to a levelthat exceeds the standard.

4.2. Contributing FactorsThe degree to which tailing and rebound

complicate remediation efforts at a site is afunction of the physical and chemical characteris-tics of the contaminant being treated, the subsur-face solids, and the ground water. Major factorsand processes that contribute to tailing andrebound are discussed below.

4.2.1. Non-Aqueous Phase LiquidsAlthough immiscible LNAPLs and DNAPLs

tend to be relatively insoluble in water, unfortu-nately they often are sufficiently soluble to causeconcentrations in ground water to exceed MCLs.Consequently, residual and pooled free-productNAPL will continue to contaminate ground waterthat makes sufficient contact to dissolve smallamounts from the NAPL surface (Figure 11a).When ground water is moving slowly, contami-nant concentrations can approach the solubilitylimit for the NAPL (Figure 11c). Although pump-and-treat systems increase ground-water velocity,causing an initial decrease in concentration, thedecline in concentration will later tail off until theNAPL’s rate of dissolution is in equilibrium withthe velocity of the pumped ground water. Ifpumping stops, the ground-water velocity slowsand concentrations can rebound, rapidly at firstand then gradually reaching the equilibrium

20

Figure 10. Concentration versus pumping duration or volume showing tailing and rebound effects (Cohen et al.,1994).

concentration (Figure 11c), unless pumping isresumed.

As shown in Table 1, DNAPL contamination inheterogeneous and fractured aquifers is the mostintractable. The reasons for this are

• DNAPLs create an unstable wetting front inthe subsurface, with fingers of more rapidvertical flow speeding the movement deeperinto the saturated zone (Figure 12). (This alsomakes accurate delineation of zones ofresidual contamination extremely difficult inhomogeneous aquifers.)

• If the volume of DNAPL exceeds the residualsaturation capacity of the unsaturated andsaturated zones, the DNAPL will reach lowerpermeability materials and form pools of freeproduct (Figure 5).

• In heterogeneous aquifers, localized lenses oflow-permeability strata may cause pools offree product to develop throughout thesaturated zone (Figure 12). Low-permeabilitystrata also may cause extensive lateral move-ment of the DNAPL. DNAPL pools areespecially problematic because the contami-nant will dissolve even more slowly thanresidual DNAPL. It may take tens of years toremove 1 cm of contaminant from a DNAPLpool (NRC, 1994).

4.2.2. Contaminant DesorptionThe movement of many organic and inorganic

contaminants in ground water is retarded bysorption processes that cause some of the dis-solved contaminant to attach to solid surfaces.The amount of contaminant sorbed is a functionof concentration, with sorption increasing asconcentrations increase, and the sorption capacity

0

Pumping On

Theoretical RemovalWithout Tailing

Removal With TailingRebound

Pumping Off

Apparent ResidualContaminantConcentration

CleanupStandard

Rel

ativ

e C

once

ntra

tion

21

Figure 11.Contaminants are mobilizedwhen ground water that isundersaturated with acontaminant comes incontact with a NAPL (a) orcontaminant sorbed on anorganic carbon or mineralsurface (b). High ground-water velocities and shortcontact times will result inlow contaminantconcentrations, and lowvelocities and long contacttimes will result in highcontaminant concentrations(c) (adapted from Gorelick etal., 1993).

Liquid-Liquid Partitioning

Advection

Desorption ofAdsorbed Contaminants

Organic Carbon orMineral Oxide Surface

Equilibrium Concentration

Dis

solv

ed C

onta

min

ant C

once

ntra

tion

Contact Time Increases to Right

Ground-Water Velocity Increases to Left

Low Ground-Water Velocities and Long ContactTimes Produce High Contaminant Concentrations(approaching equilibrium) in Ground Water

High Ground-Water Velocities and Short Contact TimesProduce Low Contaminant Concentrations in Ground Water

Contaminant Free Product

Advection

Solid Grain

(a)

(b)

(c)

22

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

23

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

of the subsurface materials. Sorbed contaminantstend to concentrate on organic matter and clay-sized mineral oxide surfaces (Figure 11b). Sorp-tion is a reversible process, however. Thus, asdissolved contaminant concentrations are reducedby pump-and-treat system operation, contami-nants sorbed to subsurface media can desorb fromthe matrix into ground water. Contaminantconcentrations resulting from sorption anddesorption show a relationship to ground-watervelocity and contact time similar to that of NAPLs(Figure 11c), causing the tailing of contaminant

concentrations during pumping as well as reboundafter pumping stops.

4.2.3. Precipitate DissolutionAs with sorption-desorption reactions, precipi-

tation-dissolution reactions are reversible. Thus,large quantities of inorganic contaminants, suchas chromate in BaCrO

4, may be found with

crystalline or amorphous precipitates in thesubsurface (Palmer and Fish, 1992). Figure 13illustrates a tailing curve where the contaminantconcentration is controlled by solubility. In thissituation, if pumping stops before the solid phaseis depleted, rebound can occur.

Contaminant ConcentrationControlled by Solubility

Solid-PhaseReserve Depleted

0

0.2

0.4

0.6

0.8

1

1.2

Pumping Duration or Volume Pumped

Rel

ativ

e D

isso

lved

Con

cent

ratio

n

24

4.2.4. Matrix DiffusionAs contaminants advance through relatively

permeable pathways in heterogeneous media,concentration gradients cause diffusion of con-taminant mass into the less permeable media(Gillham et al., 1984). Matrix diffusion is mostlikely to occur with dissolved contaminants thatare not strongly sorbed, such as inorganic anionsand some organic chemicals. During a pump-and-treat operation, dissolved contaminant concentra-tions in the relatively permeable zones are re-duced by advective flushing, causing a reversal inthe initial concentration gradient and slow diffu-sion of contaminants from the low to high perme-ability media. Figure 14, based on theoreticalcalculations of TCE concentrations in clay lensesof varying thickness, shows that diffusion is aslow process. For example, the figure indicates

that the time required to reduce the concentrationof TCE to 10 percent of the initial concentrationwould be 6 years for a clay lens 1 foot thick, 25years for a clay lens 2 feet thick, and 100 yearsfor a clay lens 4 feet thick. The significance ofmatrix diffusion increases as the length of timebetween contamination and cleanup increases. Inheterogeneous aquifers, matrix diffusion contribu-tions to tailing and rebound can be expected, aslong as contaminants have been diffusing intoless-permeable materials.

4.2.5. Ground-Water Velocity VariationTailing and rebound also result from the vari-

able travel times associated with different flowpaths taken by contaminants to an extraction well(Figure 15a-c). Ground water at the edge of acapture zone created by a pumping well travels agreater distance under a lower hydraulic gradient

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

0

0.6 Meter (2 ft)

Clay Lens Thickness = 1.2 Meters (4 ft)

Time (yr)

Ave

rage

Rel

ativ

e C

once

ntra

tion

0

20 40 60 80

0.3Meter(1 ft)

0.2

0.4

0.6

0.8

100

25

Figure 15.Tailing resulting fromground-water velocityvariations: (a) horizontalvariations in the velocity ofground water movingtoward a pumping well(Keely, 1989) lead to (b)tailing as higherconcentrations of groundwater in slower pathlinesmix with lowerconcentrations in fasterpathlines (Palmer and Fish,1992); (c) in a stratifiedsand and gravel aquifer,tailing occurs at t1 whenclean water from the uppergravel strata mixes with still-contaminated ground waterin the lower sand strata(Cohen et al., 1994).

0

0.2

0.4

0.6

0.8

1.0

1.2

Pumping Duration or Volume Pumped

Tailing Due toDifferent TravelTimes Along FlowPaths toRecovery Well

Rel

ativ

e D

isso

lved

Con

cent

ratio

n

Moderate

Fast Moderate

t0 t1

t0

t1

Stratified Sand-Gravel Aquifer

(a)

(b)

(c)

Slow

26

than ground water closer to the center of thecapture zone (Figure 15a). Additionally, contami-nant-to-well travel time varies as a function of thehydraulic conductivity in heterogeneous aquifers(Figure 15c).

4.3. Assessing the Significance ofTailing and Rebound at a Site

Determining realistic objectives for a pump-and-treat system requires sufficient site character-ization to define the complexity of the hydrogeo-logic setting (Sidebar 2) and the subsurfacedistribution of contaminants. Such informationmakes it possible for the system operator to assesswhether conditions at the site will result in tailingand rebound and to evaluate the extent to whichthese conditions are likely to increase the timeneeded to attain health-based cleanup standards.The sorption characteristics of contaminants canbe assessed using batch sorption tests with aquifermaterials (Roy et al., 1992), although aquiferheterogeneity increases the difficulty of interpret-ing test results. For organics, the potential effectsof sorption can be assessed based on a literaturereview of contaminant properties and on site-specific data on organic carbon in aquifer materi-als (Piwoni and Keely, 1990). Geochemical

computer codes can be used to assess the potentialfor tailing and rebound effects from precipitation-dissolution reactions.

Assessing the potential for removal or contain-ment of free product may be the first priority atNAPL-contaminated sites, followed by assess-ment of the extent of residual NAPL contamina-tion. For DNAPLs, residual saturation may extendthroughout the unsaturated and saturated zones(Figure 5). Typically, for LNAPLs most residualcontamination is located in the vadose zone, but itmay also extend to the depth of the seasonal lowwater table. As Figure 16 shows, pumping toremove free LNAPL product can cause residualNAPL to move deeper into the saturated zone.Consequently, when removing free-productLNAPL that is floating on the water table, stepsshould be taken to avoid or minimize movementof residual NAPL deeper into the saturated zone.

Berglund and Cvetkovic (1995) evaluated therelative importance of the degree of heterogeneityin hydraulic conductivity and mass transferprocesses and concluded that the rate of masstransfer and the extent to which contaminants aresorbed on aquifer solids are the most importantparameters that affect predicted cleanup time.

27

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

28

5. Effective Hydraulic ContainmentHydraulic containment is a design objective of

nearly all pump-and-treat systems. Where restora-tion of an aquifer to health-based standards is theoverall objective, the primary goal of containmentmust be to prevent farther spread of the contami-nant plume during restoration efforts. WhereNAPLs are present, containment using hydraulicand physical barriers might be the primaryobjective for cleanup efforts in the portion of theaquifer contaminated by free product and residualNAPL (Figures 1c and 17). In such situations aconventional pump-and-treat system might beused to restore the dissolved contaminant plume(Figure 17).

Effective hydraulic containment using pumpingwells requires the creation of horizontal andvertical capture zones that draw all contaminatedground water to the wells (Section 5.1.1) or otherhydraulic barriers (Sections 5.1.2 and 5.1.3).Failure to take aquifer anisotropy into account(Section 5.2.1) or limitations in the ability tocreate sufficient drawdown to establish capturezones (Section 5.2.2) may allow contaminants toescape from these systems. Additionally, stagna-tion zones created by pumping operations or theuse of injection wells can reduce the effectivenessof cleanup efforts (Section 5.2.3). The monitoringof both hydraulic heads (Section 6.4.1) andground-water quality (Section 6.4.2) can provideearly indications that contaminants are not beingcontained.

5.1. Ground-Water Barriers and FlowControl

Hydraulic containment can be accomplished bycontrolling the direction of ground-water flowwith capture zones (Section 5.1.1) or pressureridges (Section 5.1.2) or by using physicalbarriers (Section 5.1.3). Figure 17 illustrates apump-and-treat system that uses all three types ofhydraulic controls: (1) the contaminant sourcearea is surrounded by a barrier wall, (2) extractionwells around the margins of the dissolved plumecapture the contaminated ground water, and (3)treated ground water is reinjected to create apressure ridge along the axis of the contaminantplume. Note that the pressure ridge in Figure 17serves the function of increasing pore-volumeexchange rates rather than functioning as abarrier. Barrier pressure ridge systems are createdby placing injection wells along the perimeter of acontaminant plume.

5.1.1. Horizontal and Vertical CaptureZones

Pumping wells provide hydraulic containmentby creating a point of low hydraulic head to whichnearby ground water flows. The portion of anaquifer where flow directions are toward apumping well is called a capture zone. In anisotropic aquifer, where hydraulic conductivity isthe same in all directions, ground-water flow isperpendicular to the hydraulic head contours, alsocalled equipotential lines (Figure 18b).

29

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

A pumping well creates a zone of influencewhere the potentiometric surface has been modi-fied (Figure 18c). The capture zone is the portionof the zone of influence where ground water flowsto the pumping well (Figure 18d). Figure 15ashows how a capture zone creates flow lines ofvarying velocity. The size and shape of a capturezone depend on the interaction of numerousfactors, such as

• The hydraulic gradient and hydraulic conduc-tivity of the aquifer.

• The extent to which the aquifer is heteroge-neous (Sidebar 2) or anisotropic (Section5.2.1).

• Whether the aquifer is confined or uncon-fined.

• The pumping rate and whether other pumpingwells are operating.

• Whether the screened interval of the wellfully or partially penetrates the aquifer.

When the screened portion of a pumping wellfully penetrates an aquifer (Figure 1b), a two-dimensional analysis to delineate the horizontal

Initial Ground-Water Flow Direction

Injection Well

Extraction Well

Barrier Wall

Contaminant Source Area

Limit of Dissolved Plume

30

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

pw

(d)

Capture Zone

(b)

(c)

360

360

370

380

390

400

(a)

360

370

380

390

400

Zone of Influence

31

capture zone is usually sufficient. When a pump-ing well only partially penetrates an aquifer,however, vertical capture zone analysis also isrequired to determine whether the capture zonewill contain a contaminant plume. Figure 19shows a vertical capture zone for a partiallypenetrating well. If the contaminant plumeextended to the base of the aquifer, some contami-nants would bypass the well, despite the presenceof apparent upward gradients. In stratified aniso-tropic media (Section 5.2.1), the vertical hydrauliccontrol exerted by a partially penetrating well willbe further diminished.

5.1.2. Pressure Ridge SystemsPressure ridge systems are produced by inject-

ing uncontaminated water into the subsurfacethrough a line of injection wells locatedupgradient or downgradient of a contaminationplume. The primary purpose of a pressure ridge isto increase the hydraulic gradient and hence thevelocity of clean ground water moving into theplume, thereby increasing flow to the recoverywells, which serves to wash the aquifer.Upgradient pressure ridges also serve to divert theflow of uncontaminated ground water around theplume, and downgradient pressure ridges preventfurther expansion of the contaminant plume.Typically, treated ground water from extraction

wells within a contaminant plume supply theupgradient or downgradient injection wells usedto create a pressure ridge.

5.1.3. Physical BarriersPhysical barriers are constructed of low-

permeability material and serve to keep freshground water from entering a contaminatedaquifer zone. They also help prevent existingareas of contaminant from moving into an area ofclean ground water or releasing additional con-taminants to a dissolved contaminant plume. Mostsystems involving physical barriers also requireground-water extraction to ensure containment bymaintaining a hydraulic gradient toward thecontained area (see Figure 1c). The advantage ofphysical barriers is that the amount of groundwater that must be extracted is greatly reducedcompared to the amount when using hydrody-namic controls, as described in Sections 5.1.1 and5.1.2. Major types of barriers include

• Caps (or covers), which are made of low-permeability material at the ground surface,can be constructed of native soils, clays,synthetic membranes, soil cement, bitumi-nous concrete, or asphalt.

• Slurry trench walls, excavated at the properlocation and to the desired depth while

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

984

982

980

978 976

974

Vertical CaptureZone

970 968972

32

keeping the trench filled with a clay slurry,keep the trench sidewalls from collapsing andbackfilling with soil bentonite, cementbentonite, or concrete mixtures.

• Grout curtains are created by injectingstabilizing materials under pressure into thesubsurface to fill voids, cracks, fissures, orother openings in the subsurface. Grout alsocan be mixed with soil using larger augers.

• Sheet piling cutoff walls are constructed bydriving sheet materials, usually steel, throughunconsolidated materials with a pile driver ormore specialized vibratory drivers.

Knox et al. (1984) provide further informationon the design and construction of physicalground-water barriers.

5.2. Hydraulic Containment: OtherSpecial Considerations

Certain site conditions can allow contaminantsto escape from a hydraulic containment system ifthey are not characterized and anticipated.

5.2.1. Effects of AnisotropyIn anisotropic aquifers, hydraulic conductivity

varies with direction. In flat-lying sedimentaryaquifers, hydraulic conductivity is often higher ina horizontal than a vertical direction. In fracturedrock and foliated metamorphic rocks, such asschist, the direction of maximum and minimumpermeability is usually aligned parallel andperpendicular, respectively, to foliation or beddingplane fractures (Cohen et al., 1994). Wheresedimentary strata and foliated media are inclinedor dipping, significant horizontal anisotropy maybe an aquifer characteristic. In anisotropic media,the flow of ground water, as well as contaminantsmoving with ground water, is usually not perpen-dicular to the hydraulic gradient.

Figure 20 illustrates how horizontal anisotropyin fractured rock can change the location of the

capture zone of a pumping well. In an aquifer thatis assumed to be isotropic, the general direction ofground-water flow should be perpendicular to thehydraulic gradient (Figure 20a). If fractures causehydraulic conductivity to be higher in a north-south rather than an east-west direction, however,the direction of ground-water flow will divergefrom the direction of the hydraulic gradient(Figure 20b). In this example, siting a pumpingwell based only on the hydraulic gradient (Figure20a) would result in its failure to capture anyportion of a contaminant plume, except in theimmediate vicinity of the well.

A contaminant plume that does not follow thehydraulic gradient may indicate that anisotropy isinfluencing the direction of ground-water flow.Aquifer heterogeneities, such as buried streamchannels that have a different direction than thehydraulic gradient, also may allow the directionof contaminant travel to diverge from the hydrau-lic gradient. Computer programs, such as EPA’sWell Head Protection Area (WHPA) code, can beuseful for evaluating the potential effects ofanisotropy on well capture zones. Figure 21shows such a simulation for three pumping wells.In this case, with a vertical to horizontal anisot-ropy ratio of 10:1, the orientation of the capturezones shifts from northwest-southeast (isotropic)to east-west (anisotropic).

5.2.2. Drawdown LimitationsUnder some conditions creating and maintain-

ing an inward hydraulic gradient for a contami-nant plume is problematic. In such situations,injection wells may be required to create pressureridges (Section 5.1.2) or physical barriers mayneed to be installed (Section 5.1.3). Site condi-tions that might indicate the need for such mea-sures include (Cohen et al., 1994)

• Limited saturated thickness of the aquifer

• Relatively high initial hydraulic gradient

33

• Sloping aquifer base

• Very high aquifer permeability

• Low aquifer permeability

Where these conditions exist and hydrauliccontainment is planned, particular care should betaken during site characterization and pilot tests toassess drawdown limitations.

5.2.3. Stagnation ZonesStagnation zones develop in areas where pump-

and-treat operations create low hydraulic gradi-ents and, consequently, low ground-water veloci-ties. The stagnation zone associated with a singleextraction well is likely to be locateddowngradient from the well (Figure 22a). A

stagnation zone can develop upgradient from aninjection well, however, and form in low-perme-ability zones, regardless of hydraulic gradient.When multiple extraction or injection wells areinvolved, a number of stagnation zones maydevelop (Figure 22b). Stagnation zones caused bylow hydraulic gradients can be identified bymeasuring hydraulic gradients, tracer movement,and ground-water flow rates using downholeflowmeters and through modeling analysis.Stagnation zones within a contaminant plume canreduce the efficiency of a pump-and-treat system;thus, minimizing stagnation is an importantobjective of capture zone analysis and optimiza-tion modeling (Section 6.1).

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

K K K

K

Y Y X

X

=

Zoneof Contribution

General

Direction of

Ground-W

ater Flow

and Hydraulic

Gradient

PumpingWellWater-

TableContours

200

175

150

125

Isotropic Aquifer Anisotropic Aquifer(a) (b)

Water-Table Contours

Zone ofContribution

GeneralDirection of

Hydraulic

Gradient

General D

irection of

Ground-W

ater Flow

PumpingWell

K

KK

K

Y

Y X

X

=5

200

175

150

125

34

Figure 21.Capture zonesimulation ofthree pumpingwells for anisotropic aquifer(a) andanisotropy ratio of10:1 (b) using theEPA WHPA code.

9000

7200

5400

3600

1800

00 2100 4200 6300 8400 10500

(ft)

(ft)

(ft)9000

7200

5400

1800

3600

00 2100 4200 6300 8400 10500

(ft)

Anisotropic

Isotropic

(a)

(b)

35

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

531

532

533

534

StagnationPoint

1 2 3 4

5

5

6

6

891011

8910

11

12

13

15

20

12

(a)

535 536538

537

535538

537

535

536

538

537

535

536

538537

539

540

541542

543

InjectionWell

Injection Well

345678

4567

8

5 6 7 8

345

67

810

6

6

7

6

7

67 8

810

10

15 2030

(b)

43 10

547

546

545

544

543

542

541

540 53

9

535

538 53

4

PumpWell

Hydraulic HeadContour Map Ground-Water Velocity

Contour Map

10

3

Ground-Water Velocity Contour Map

537

536

536

Extraction Well Extraction Well

Extraction Well Extraction Well

Hydraulic HeadContour Map

36

The basic operating principle of a pump-and-treat system calls for locating a well (or wells)and then pumping at rates that cause all water in acontaminant plume to enter the well rather thancontinue traveling through the subsurface. Table 2lists types of data required for evaluating thefeasibility of using the pump-and-treat approachat a contaminated ground-water site and thendesigning an appropriate system. This sectiondescribes the key aspects of designing andoperating a pump-and-treat system for optimalperformance.

6.1. Capture Zone Analysis andOptimization Modeling

In recent years, numerous mathematical modelshave been developed or applied to computecapture zone, ground-water pathlines, and associ-ated travel times to extraction wells or drains. Forrelatively simple hydrogeologic settings (homoge-neous isotropic aquifers), analytical equationssolved manually, using graphical techniques orcomputer codes based on analytical solutions,may be adequate. For more complex sites, nu-merical computer models may be required. Thesemodels provide insight to flow patterns generatedby alternative pump-and-treat approaches and tothe selection of monitoring points and frequency.The WHPA model (Blandford and Huyakorn,1991) and Capture Zone Analytic Element Model(CZAEM) (Haitjema et al., 1994; Strack et al.,1994) developed by EPA are examples of rela-tively simple computer software based on analyti-

cal equations (WHPA) and the innovative analyticelement method (CZAEM) that allows capturezone and ground-water pathline analysis. Thenumerical MODFLOW and MODPATH modelsdeveloped by the U.S. Geological Survey arecommonly used to model more complex hydro-geologic settings. Cohen et al. (1994) identify anumber of computer codes of potential value forcapture zone analysis. More detailed informationabout specific models and EPA guidance on theuse of models are available in references onGround-Water Modeling at the end of this guide(Section 9). Sidebar 5 summarizes the results ofcomputer modeling performed to evaluate theeffect of different hydrogeologic conditions on theeffectiveness of different types of well patterns.

In addition, optimization programming methodsare being used increasingly to improve pump-and-treat system design (Gorelick et al., 1993). Asapplied to the design of pumping systems, optimi-zation involves defining an objective function,such as minimizing the sum of pumping ratesfrom a number of wells. A set of restrictions, orconstraints, specify various conditions, such asmaximum pumping rates and minimum hydraulicheads at individual wells, that must be satisfied bythe optimal solution alternative. Hydrauliccontainment of a contaminant plume usuallyrequires only linear optimization methods, butwhen contaminant concentrations are specified asconstraints, nonlinear methods are often required(Rogers et al., 1995). At the Lawrence Livermore

6. Pump-and-Treat System Design and Operation

37

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

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

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

approach

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

applicability and evaluateeffectiveness

Contaminant/soil properties To determine mobility Published literature, laboratory(density, aqueous solubility, properties; applicability of testsoctanol-water/carbon pump-and-treat approachpartitioning coefficient, soilorganic carbon content,sorption parameters)

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

Depth to aquifer/water table To select appropriate extraction Hydrogeologic maps,system type; consideration for observation wells, boring logs,implementation piezometers

Ground-water flow direction To determine proper well Water level data,and vertical/horizontal gradients locations/spacing considerations potentiometric maps

for implementation

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

implementation

NAPL density/viscosity/ To predict vertical distribution Literature, laboratorysolubility; residual saturation of of contamination; consideration measurementsvadose zone and saturated zone for implementation and

evaluating effectiveness

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

Precipitation/recharge To calculate water balance; NOAA reports, local weatherconsideration for implementing bureaus; onsite measurementsand evaluating effectiveness

Locations, screen/open interval To determine Well inventory, pumpagedepths, and pumping rates of impacts/interference; recordswells influenced by site considerations for implementing

and evaluating effectiveness

38

Sidebar 5Computer Modeling of Well Patterns Versus Hydrogeologic Conditions

Satkin and Bedient (1988) used the U.S. GeologicalSurvey MOC model to evaluate the effectiveness ofseven different well patterns (Figure 23) for restoring

contaminated ground water under eight generichydrogeologic conditions. The hydrogeologic settingswere defined as various combinations of three majorfactors: maximum drawdown (high > 10 ft; low < 5 ft),hydraulic gradient (high = 0.008; low = 0.0008), andlongitudinal dispersivity (high = 30 ft; low = 10 ft).

Because the contaminants were assumed to not interactwith aquifer solids, tailing and rebound effects were nota consideration in the study. Major conclusions of thecomputer simulations include the following:

• Significant differences in cleanup time were observedusing various well locations for a given well pattern.

• The three-spot, doublet, and double-cell wellpatterns are effective under low hydraulic gradientconditions. These well patterns minimize cleanup

time, volume of water circulated, and volume ofwater treated.

• The three-spot well pattern performed better thanany of the other well patterns studied under a highhydraulic gradient, high drawdown, and either a lowor high dispersivity.

• None of the well patterns investigated was able tocontain and clean up the contaminated plume in asetting with high gradient, low drawdown, and highdispersivity.

• The centerline well pattern is effective in achievingup to 99 percent contaminant reduction under bothlow and high gradient conditions, but it may present

a water disposal problem.

• The five-spot well pattern was the least effective ofthe well patterns studied.

39

x

x

x

x

Single Doublet Centerline

3-Spot

5-Spot Double Triangle Double Cell

..Pumping Well x ..Injection Well

xx

xx

x

x x

x

Figure 23.Major types ofpumping/injectionwell patterns(Satkin andBedient, 1988).

40

National Laboratory (LLNL) site, Rogers et al.(1995) applied an innovative nonlinear optimiza-tion approach, using artificial neural networks anda genetic algorithm, to evaluate more than 4million pumping patterns for the project’s 28extraction and injection wells. The three top-ranked patterns required 8 to 13 wells, withprojected costs estimated at $41 to $53 millionover the 50-year project life. Using these pumpingpatterns was estimated to cost from one-third toone-quarter the cost of using all 28 wells at anestimated cost of $155 million.

6.2. Efficient Pumping OperationsRemoval of contaminated ground water should

be a dynamic process that uses information on theresponse of the ground-water system to improvethe efficiency of pumping operations (Section3.3). Elements of efficient pumping operationscan include

• Combined plume containment and sourceremediation, which can be achieved throughthe design of the initial pumping flow field.For example, at the LLNL site a line ofextraction wells at the downgradient marginsof the plume were established to prevent themovement of contaminants toward municipalwater-supply wells, while other wells werelocated in the source areas where the contami-nant concentrations were highest. This limitedthe area requiring remediation and maximizedcontaminant removal (Hoffman, 1993).

• Phased construction of extraction wells,which allows data on the monitored responseof the aquifer to pumping operations to beused in siting subsequent wells.

• Adaptive pumping, which involves designingthe well field such that extraction and injec-tion can be varied to reduce zones of stagna-tion. Extraction wells can be periodically shut

off, others turned on, and pumping ratesvaried to ensure that contaminant plumes areremediated at the fastest rate possible. Figure24 illustrates stagnation zones that woulddevelop at the LLNL site if a fixed pumpingwell configuration were used. With thisapproach, remediation at the site would takeabout 100 years (Figure 25). Computermodeling of adaptive pumping indicates thatthis technique should make it possible toreduce the time required for site cleanup toabout 50 years (Figure 25). Further refine-ments in design might shorten the time evenfurther (Hoffman, 1993).

• Pulsed pumping, which has the potential toincrease the ratio of contaminant massremoved to ground-water volume where masstransfer limitations restrict dissolved contami-nant concentrations. Figure 26 illustrates theconcept of pulsed pumping. During theresting phase of pulse pumping, contaminantconcentrations increase due to diffusion,desorption, and dissolution in slower movingground water (Figure 11). Once pumping isresumed, ground water with a higher concen-tration of contaminants is removed, thusincreasing mass removal during pumping.Special care must be taken to ensure that thehydraulic containment objective is met duringpump rest periods. Bartow and Davenport(1995) have reported that about 19 percent ofthe pump-and-treat systems in Santa ClaraValley, California, use some form of pulsedpumping. A recent study by Harvey et al.(1994), however, on the effects of physicalparameters (e.g., the mass transfer ratecoefficient) concluded that pulsed pumpingprovides little if any advantage over continu-ous pumping at an average rate.

41

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

42

Figure 25.Effect of adaptivepumping oncleanup time atLawrenceLivermoreNationalLaboratorySuperfund site(Cohen et al.,1994, afterHoffman, 1993).

6.3. Treating Contaminated GroundWater

Once extraction wells have brought contami-nated water to the surface, treatment is relativelystraightforward, provided that appropriate meth-ods have been selected and the capacity of thetreatment facility is adequate. Table 3 summarizesthe applicability of various treatment technologiesto ground water contaminated by any of the majorcategories of inorganic and organic contaminants.U.S. EPA (1995) describes conventional technolo-gies that have evolved from industrial wastewatertreatment and that have been implemented at fullscale for treatment of contaminated ground water.These methods fall into two main categories:

• Biological. Biological treatment methods usemicroorganisms to degrade organic com-pounds and materials into inorganic products.The methods may be applicable for treatmentof ground water contaminated by organiccompounds if concentrations are low enoughand the biological processes are not inhibited.The best established biological treatmentmethods include (1) activated sludge systems,(2) a sequencing batch reactor, (3) powderedactivated carbon in activated sludge (bio-physical system), (4) rotating biologicalcontactors, and (5) an aerobic fluidized bedbiological reactor.

10

100

1000

0 5 10 15 20 25 30 35 40 45 50Time (yr)

Hig

hest

Res

idua

l VO

C C

once

ntra

tion

Fixed WellConfiguration

AdaptivePumping

43

Figure 26.The pulsedpumping concept(Cohen et al., 1994,after Keely, 1989).

• Physical/Chemical. Physical, chemical, or acombination of physical and chemical meth-ods can be used to remove contaminants fromground water. The most commonly usedmethods include (1) air stripping, (2) acti-vated carbon, (3) ion exchange, (4) reverseosmosis, (5) chemical precipitation of metals,(6) chemical oxidation, (7) chemicallyassisted clarification, (8) filtration, and (9)ultraviolet (UV) radiation oxidation.

Various emerging and innovative treatmenttechnologies, such as electrochemical separationand wet air oxidation, are being tested. The EPAreference sources identified for ground-watertreatment methods at the end of this guide (Sec-tion 9) provide additional information on estab-lished and innovative treatment technologies.

6.4. Monitoring PerformanceAn appropriately designed monitoring program

is essential for measuring the effectiveness of apump-and-treat system in meeting hydrauliccontainment and aquifer restoration objectives. Ingeneral, containment monitoring involves (1)

measuring hydraulic heads to determine if thepump-and-treat system creates inward gradientsthat prevent ground-water flow and dissolvedcontaminant migration across the containmentzone boundary, and (2) ground-water qualitymonitoring to detect any contaminant movementor increase of contaminant mass across thecontainment zone boundary. Aquifer restorationmonitoring mainly involves measurement ofcontaminant concentrations in pumping andobservation wells to determine the rate andeffectiveness of mass removal. Cohen et al.(1994) provide more detailed guidance on moni-toring the performance of pump-and-treat sys-tems.

6.4.1. Hydraulic Head Monitoring forContainment

In general, the number of observation wellsneeded for monitoring inward hydraulic gradientsin a containment area increases with site complex-ity and with decreasing gradients along thecontainment perimeter. Strategies for adequatelymonitoring inward gradients and hydrauliccontainment include (Cohen et al., 1994)

ResidualContamination

1 2 3 4 5 6 7 8

Time

Con

cent

ratio

nP

ump

On

Off

Max

0

t t t t t t t t

44

• Measuring hydraulic heads in three dimen-sions using nested piezometers for detectingvertical gradients. As shown in Figure 19,partially penetrating wells may not create anadequate vertical capture zone. Where leakyconfining layers separate aquifers, hydraulicgradients should be toward the contaminatedzone. Figure 27 illustrates observations froma nest of piezometers at the Chem-DyneSuperfund site in Ohio. Water levels from thedeep piezometer are consistently about a foothigher than in the intermediate and shallowpiezometers, indicating an upward gradient.

• Monitoring water levels in observation wellsintensively during system startup and equili-bration to determine an appropriate measure-ment frequency. This may involve usingpressure transducers and dataloggers to makenear-continuous head measurements for a fewdays or weeks, then switching sequentially todaily, weekly, monthly, and possibly quarterlymonitoring. Data collected during each phaseshould provide the justification for anysubsequent decrease in monitoring frequency.

• Making relatively frequent hydraulic headmeasurements when the pumping rates orlocations are modified, or when the system issignificantly perturbed in a manner that hasnot been evaluated previously. Significantnew perturbations can arise from, for ex-ample, unusual recharge, flooding, drought,and new offsite well pumping.

• Measuring hydraulic head as close to thesame time as possible when monitoringinward hydraulic gradients or a potentiomet-ric surface so that data are temporally consis-tent. This ensures that differences in ground-water elevation within a network representspatial rather than temporal variations.

• Supplementing hydraulic head data with flow-path analysis using potentiometric maps orparticle tracking computer codes. Figure 28shows that ground water can flow betweenand beyond recovery wells even thoughhydraulic heads throughout the mappedaquifer are higher than the pumping level.

• Conducting an analysis to determine ifcontainment is threatened or lost whenhydraulic head data do not indicate a clearinward gradient. Rose diagrams can beprepared to display the variation over time ofhydraulic gradient direction and magnitudebased on data from at least three wells (Figure29). Even when the time-averaged flow istoward the pump-and-treat system, contain-ment can be compromised if contaminantescapes from the larger capture zone duringtransient events or if a net component ofmigration away from the pumping wellsoccurs over time.

6.4.2. Ground-Water Quality Monitoringfor Containment

Monitor well locations and completion depthsshould be selected to provide a high probability ofdetecting containment system leaks in a timelymanner. Consequently, monitor wells withrelatively close spacing are usually located alongor near the potential downgradient containmentboundary. Ground-water quality sampling usuallyis performed less frequently than the measuring ofhydraulic head because contaminant movement isa slower process. Because ground-water qualitymonitoring is more expensive than hydraulic headmonitoring, designing a cost-effective monitoringplan requires special care. Strategies that mayhelp reduce costs without compromising theintegrity of the program include (Cohen et al.,1994)

45

Metals

Heavy metals X ● ● X X ❍ X X X ❍ ● ● X ● ● ● X ●

Hexavalent chromium X ● X X X ● X X X ❍ ● X X ❍ ● X X ●

Arsenic X ❍ ● ❍ ❍ X X X X ❍ X ❍ X ● ● ● X X

Mercury X ● ● X X ● X X X ● X ❍ X ❍ ● ● X X

Cyanide X X X ● ● X X X X X ● X X ● ● X ❍ ❍

Corrosives ● ● X X X X ❍ X X X X X X X X X X X

Volatile organics X X X ❍ ● X ● ● ● ● X X X ❍ ❍ X ❍ X

Ketones X X X ❍ ● X ● ● ● X X X X X X X ● X

Semivolatile organics X ❍ ❍ ● ● X ● X ● ● ❍ ❍ ❍ ● ● X ● X

Pesticides X ❍ ❍ ● ● X ● X ❍ ● ❍ ❍ ❍ ● ● ● ❍ X

PCBs X ● ● ● ● X ● X X ● ● ● ● ● ● ● ❍ X

Dioxins X ● ● ● ❍ X ● X X ● ● ● ● ● ● ● ❍ X

Oil and grease/floating X ● ● X X X ● X X X ● ● ● ● ● ❍ ❍ Xproducts

● Applicable ❍ Potentially Applicable X Not Applicable

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

Neu

tral

izat

ion

Pre

cipi

tatio

n

Cop

reci

pita

tion/

Coa

gula

tion

UV

/Ozo

ne

Che

mic

al O

xida

tion

Red

uctio

n

Dis

tilla

tion

Air

Str

ippi

ng

Ste

am S

trip

ping

Act

ivat

ed C

arbo

n

Eva

pora

tion

Gra

vity

Sep

arat

ion

Flo

tatio

n

Mem

bran

e S

epar

atio

n*

Ion

Exc

hang

e

Filt

ratio

n

Bio

logi

cal

Ele

ctro

chem

ical

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

Contaminants

46

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

568

567

566

565

564

563

562

Wat

er-L

evel

Ele

vatio

n (f

t)

565.36

564.34

564.24

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

Water-Level Measurement in Deep Piezometer

Water-Level Measurement in Intermediate Piezometer

Water-Level Measurement in Shallow Piezometer

Daily Average Water Level From Hourly Data Recordedin Shallow Piezometer

Time-Averaged Water Level in Deep Piezometer

Time-Averaged Water Level in Intermediate Piezometer

Time-Averaged Water Level in Shallow Piezometer

1992

Explanation

47

• Sampling more frequently and performingmore detailed chemical analyses in the earlyphase of the monitoring program, and usingthe information gained to optimize samplingefficiency and reduce the spatial density andtemporal frequency of sampling in the laterphases.

• Monitoring ground-water quality in perimeterand near-perimeter leak detection wells morefrequently than in wells that are at a greaterdistance from the contaminant plume limit.

• Specifying sampling frequency based onpotential containment failure migration ratesthat factor in hydraulic conductivity andeffective porosity of the different media as

well as the maximum plausible outwardhydraulic gradients. Consider more frequentsampling of more permeable strata in whichmigration might occur relatively quickly ascompared to the sampling frequency for lesspermeable media.

• Focusing chemical analyses on site contami-nants of concern and indicator constituentsafter performing detailed chemical analysesduring the remedial investigation or the earlyphase of a monitoring program. Conductmore detailed chemical analyses less fre-quently or when justified based on the resultsof the more limited analyses.

Figure 28.Ground-waterflow between andbeyond theextraction wells,resulting eventhough hydraulicheads throughoutthe mappedaquifer are higherthan the pumpinglevel (Cohen etal., 1994).

362.0

362.

5

362.5Constant-Head

Pumping

Elevation in

Each Well Is

360.0 ft

363.0362.5

363.5

363.0

364.0

363.5

362.0

48

6.4.3. Aquifer Restoration MonitoringAquifer restoration monitoring consists of three

main elements:

• Ground-water sampling from all extractionwells and selected observation wells withinthe contaminant plume to interpret cleanupprogress. Parameters analyzed should include(1) the chemicals of concern, (2) chemicalsthat could affect the treatment system, such asiron, which can precipitate and clog treatmentunits if ground water is aerated, and (3)

chemicals that may indicate the occurrence ofother processes of interest, such as dissolvedoxygen, carbon dioxide, and biodegradationproducts. These sampling data are importantfor making adjustments for efficient welloperation (Section 6.2).

• Periodic sampling and chemical analysis ofaquifer materials from representative loca-tions in the contamination zone to measureremoval of nondissolved contaminants.

Figure 29.Example display of ground-waterflow directions and hydraulicgradients determined betweenthree observation wells (Cohen etal., 1994).

0.00

10.00

20.00

30.00

40.00

50.00

6

West

North

East

South

49

• Regular sampling and analysis of treatmentsystem influent and effluent to assess (1)treatment system performance, (2) change ininfluent chemistry that may influence treat-ment effectiveness, and (3) dissolved con-taminant concentration trends. Figure 30shows influent and effluent VOC concentra-tions for the first 6 years of operation at theChem-Dyne Superfund site treatment plant.Influent concentrations data showed a largedrop in the first year, and then a more gradualdecline over the next 5 years due to tailingeffects (Section 4).

The simplest indicator of progress in removingground-water contaminants is a plot of thecumulative mass removed from the aquifer asmeasured by influent concentrations to thetreatment system. Figure 31 shows the cumulativemass of VOC removal at the Chem-Dyne site.Approximately 27,000 pounds of VOCs have beenremoved since the system became operational. Asis apparent from both Figures 30 and 31, however,the rate of removal slowed significantly in thesixth year. Consequently, removal of the remain-ing one-third of the in-place mass will take muchlonger than 6 years.

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

400

300

200

100

0 1991 1992C

once

ntra

tion,

in µ

g/L

19921991

2

4

6

8

10

12

Con

cent

ratio

n, in

mg/

L

Influent Effluent

Year19901989198819870

50

6.5. Evaluating Restoration Successand Closure

Ground-water restoration, as operationallydefined, is achieved when a predefined cleanupstandard is attained and sustained. Figure 32outlines procedures for determining the successand/or timeliness of closure of a pump-and-treatsystem. U.S. EPA (1992) defines six stages ofremediation using water quality data from a singlewell (Figure 33):

• Stage 1. Site evaluation to determine the needfor and conditions of a remedial action; definecleanup standard.

• Stage 2. Operation of the remediation system,during which contaminant concentrationsdecline.

• Stage 3. Conclusion of treatment after con-taminant concentrations have remained belowthe cleanup standard for a sufficient period oftime based on expert knowledge of theground-water system and data collectedduring pump-and-treat operations.

• Stage 4. Post-termination monitoring of waterlevels and contaminant concentrations todetermine when the ground-water flowsystem is reestablished.

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

30,000

25,000

20,000

15,000

10,000

5,000

1987 1988 1989 1990 1991 19920

Year

Mas

s of

Prio

rity

Pol

luta

nt V

OC

s, in

lb

51

Figure 32.Determining thesuccess and/ortimeliness ofclosure of apump-and-treatsystem (Cohen etal., 1994).

Define Attainment Objectives andCleanup Standard

Develop Sampling and AnalysisPlan for Performance

Monitor SystemPerformance

Is the CleanupStandard Reached?

TerminateTreatment

Allow System to ReachSteady-State

Verify the Attainment ofCleanup Standard

Is the CleanupStandard Maintained

Over Time?

Monitor asNecessary

Assess and ReviseTreatment Design as

Necessary

ModifyRemedial

ActionObjectives

Demonstration ofTechnical

Impracticability

No No

Yes

No

Yes

Continue/Modify

Treatment

52

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

• Stage 5. Sampling to assess attainment of thecleanup standard. If the treatment standard isnot met, the treatment design may need to beassessed and revised (Figure 32).

• Stage 6. Declaration that the aquifer is cleanor still contaminated based on data collectedduring Stage 5.

Cohen et al. (1994) and U.S. EPA (1992)address in more detail the types of statisticaltechniques that are required to analyze short-termand long-term trends in contaminant concentra-tions.

0.2

0

0.4

0.6

0.8

1

1.2

Mea

sure

d G

roun

d-W

ater

Con

cent

ratio

n

CleanupStandard

1 2 3

4 5

6

StartTreatment

End SamplingDeclare Clean orContaminated

Date

EndTreatment

StartSampling

53

7. Variations and Alternatives to Conventional Pump-and-Treat MethodsNumerous variations and enhancements of

pump-and-treat systems are possible. Major typesinclude

• Using trenches or drains in combination withor to replace vertical pumping wells (Section7.1). Where site conditions are favorable (i.e.,shallow contamination), trenches are acommonly used method for interceptingcontaminated ground water.

• Using horizontal wells or trenches to replaceor complement vertical wells (Section 7.1).Recent developments in directional drillingtechnology make the use of horizontal orinclined wells an attractive alternative ap-proach.

• Inducing fractures in the subsurface toimprove the yield of wells (Section 7.1).Although widely used by the petroleumindustry, the use of induced fractures isconsidered an emerging technology inground-water remediation with applicationslimited to contaminated ground water in low-permeability materials.

• Implementing vadose zone source control andremediation, often as a necessary adjunct toground-water cleanup (Section 7.2).

• Making chemical enhancements, which canhave the potential to accelerate aquiferremediation (Section 7.3).

• Making biological enhancements, which canpresent opportunities for eliminating or

reducing the requirements for surface treat-ment of contaminated ground water (Section7.4).

Notably few alternatives to pump-and-treatsystems are without requirements for continuousenergy input for pumping fluids (Section 7.4).

7.1. Alternative Methods for FluidDelivery and Recovery

Conventional pump-and-treat systems usuallyinvolve extraction wells—and possibly injectionwells—placed vertically in an aquifer. Alternativemethods of delivery and recovery of contaminatedground water might enhance the performance of apump-and-treat system, especially while interimmeasures are undertaken, by improving theeffectiveness of containment. These methods alsomight augment the performance of a variety ofremedial actions selected as possible long-termremedies. Major alternatives include

• Interceptor Trenches. After vertical wells,trenches are the most widely used method forcontrolling subsurface fluids and recoveringcontaminants. They function similarly tohorizontal wells, but also can have a signifi-cant vertical component, which cuts acrossand can allow access to the permeable layersin interbedded sediments. For shallowerapplications, trenches can be installed atrelatively low cost using conventional equip-ment. Recent innovations combine trenchexcavation and well screen installation into a

54

single step for depths up to 20 feet (U.S. EPA,1994). Where depth is not a constraint,interceptor trenches are generally superior tovertical wells. In such situations, they areespecially effective in low-permeabilitymaterials and heterogeneous aquifers.

• Horizontal and Inclined Wells. Relativelyrecent advances in directional drilling tech-nology, which use specialized bits to curvebores in a controlled arc, have revolutionizedthe field of well design. Directional drillingmethods can create wellbores with almost anytrajectory. Wells that curve to a horizontalorientation are especially suited to environ-mental applications (Figure 34).

• Induced Fractures. EPA research has shownthat petroleum engineering technology usedto induce fractures for increased productivityof oil wells also can improve the performanceof environmental wells. Induced fractures areused mainly where low-permeability aquifermaterials create problems for the recovery ofcontaminants.

Table 4 rates the potential applications ofalternative methods for delivery or recovery ofsubsurface fluids in relation to (1) access, (2)depth, (3) recovered phases, (4) geology, and (5)availability. Figure 35 illustrates two ways inwhich horizontal wells or trenches can be used tointercept a contaminant plume. In many applica-tions, deciding between use of a trench or ahorizontal well hinges on economic rather thantechnical issues, with trenches generally beingmore cost effective at depths less than 20 feet andhorizontal wells being generally more costeffective at depths greater than 20 feet. Costsavings can be substantial compared to verticalwell systems. For example, initial remediationplans at a site in North Carolina called for 100vertical wells to recover a hydrocarbon plume at

an estimated cost of $1 million. Instead, a con-tinuous excavation and completion system wasinstalled for less than $350,000 (U.S. EPA, 1994).EPA’s Manual Alterative Methods for FluidDelivery and Recovery (U.S. EPA, 1994) providesmore detailed information on design consider-ations and applications of these methods.

7.2. Vadose Zone Source ControlRemoval of contaminants from the vadose

(unsaturated) zone is an essential part of anyremedial action plan to clean up contaminatedground water. Major methods include

• Capping to reduce infiltration of precipitation.

• Excavation to remove contaminated soil forex situ treatment, which is most commonlyused where contaminants have not penetrateddeeply into the subsurface.

• Soil vapor extraction (SVE), which is used toextract volatile organic contaminants byflushing with air, and bioventing, a SVEsystem in which the addition of nutrientsfurther enhances the biodegradation oforganic contaminants. Both techniques,considered innovative technologies a fewyears ago, are widely used.

• In situ thermal technologies to enhance themobility of volatile and semivolatile organiccontaminants; for example, steam-enhancedextraction and radio frequency heating arepromising innovative technologies.

7.3. Physical and Chemical En-hancements

Physical and chemical enhancements to pump-and-treat systems primarily function by enhancingthe mobility of contaminants, thus increasing theirrecovery in ground water that has been pumped tothe surface for treatment. Some chemical en-hancements transform contaminants in place inthe subsurface to reduce toxicity.

55

Figure 34. Some applications of horizontal wells: (a) intersecting flat-lying layers, (b) intercepting plume elongatedby regional gradient, (c) intersecting vertical fractures, and (d) access beneath structures (U.S. EPA,1994).

(a)

(b)

(c)

(d)

56

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

Issue Horizontal Well Induced Fracture Trench

Access

Fragile structures ● Minimal surface disturbance ■ Evaluate effects of surface ■ Excavation expected to beover target displacement infeasible

Poor access over ● Standoff required ◆ Possible with horizontal ■ Excavation expected to betarget well infeasible

Depth

<6m ● 1 m minimum depth ● 1-2 m minimum depth ● Installation with commonequipment

6-20 m ● Cost of guidance system ● ❀ Excavation costs increaseincreases at >6 m with depth

>20 m ● No depth limit within ● No depth limit within ◆ Specialized excavationenvironmental applications environmental applications methods required

Recovered Phase

Aqueous ● ● ●

LNAPL ◆ Requires accurate drilling; ❀ Best with access to ● Widely used to ensurebest if water table fluctuations individual fractures capture; accommodatesare minor water table fluctuations

DNAPL ❀ Requires accurate drilling and ◆ Caution; steeply dipping ● Assuming mobile phasesite characterization fractures may cause present and accurately

downward movement located

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

Geology

Normally ❀ Smearing of bore wall may ◆ Induced fractures may be ● Large discharge expectedconsolidated clay reduce performance vertical and limited in size relative to alternatives

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

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

Stratified sediment ◆ Anisotropy may limit vertical ❀ Stratification may limit ● Good way to access manyor rock influence of well upward propagation and thin beds or horizontal

increase fracture size partings

(Continued)

57

Table 4. (Continued)

Issue Horizontal Well Induced Fracture Trench

Vertically fractured ● Orient well normal to ❀ Good where induced ❀ Orient trenchsediment or rock fractures when possible fractures cross-cut natural perpendicular to natural

fractures fractures when possible(overconsolidatedsediment and rock)

Coarse gravel ◆ Possible problems with hole ■ Permeability enhancement ● Stability a concern duringstability; penetrating cobbles may be unnecessary excavation

Thick sand ❀ May be difficult to access top ■ Permeability enhancement ● Stability a concern duringand bottom of formation; hole may be unnecessary excavationstability problems

Rock ◆ Feasible, but drilling costs ● Widely used in oil, gas, ■ Excavation difficult butmore in rock than in sediment and water wells drilled in blasting possible to make

rock trench-like feature

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

trench will require mobilization

Current Experience 150 to 250 wells at 50 to 100 sites 200 to 400 fractures at 20 to 40 1,000+ trenches at many(Approximate) sites hundreds of sites

Key● Good application

❀ Moderately good◆ Fair, with possible technical difficulties■ Poor; not recommended using available methods

58

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

(a)

(b)

Plume

StreamHorizontal Well

Source

Horizontal Well

Capture Zone Source

Plume

59

7.3.1. Physical EnhancementsAir sparging, also known as in situ aeration, is

an approach that is similar to soil vapor extrac-tion except that air is injected into the saturatedzone rather than the vadose zone (Figure 36). Airsparging systems can effectively remove asubstantial amount of volatile aromatic andchlorinated hydrocarbons in a variety of geologicsettings, but significant questions remain aboutthe ability of this technology to achieve health-based standards throughout the saturated zone(NRC, 1994). Thermal enhancements, such assteam and hot-water flooding, increase themobility of volatile and semivolatile contami-nants. Use of induced fractures (Section 7.1) isanother form of physical enhancement to pump-and-treat systems.

7.3.2. Chemical EnhancementsChemically enhanced pump-and-treats systems

require use of injection wells to deliver reactiveagents to the contaminant plume and extractionwells to remove reactive agents and contaminants(Figure 37). The major types of chemical en-hancements are

• Soil flushing, which enhances recovery ofcontaminants with low water solubility, free-product and residual NAPLs, and sorbedcontaminants. Two major types of chemicalagents can be used: (1) cosolvents, which,when mixed with water, increase the solubil-ity of some organic compounds, and (2)surfactants, which may cause contaminantsto desorb and may increase NAPL mobilityby lowering the interfacial tension betweenthe NAPL and water, increasing the solubil-ity. Soil flushing is one of the most promis-ing innovative technologies for dealing withseparate phase DNAPLs in the subsurface(NRC, 1994).

• In situ chemical treatment, which involvesreactive agents that oxidize or reduce con-taminants, converting them to nontoxic formsor immobilizing them to minimize contami-nant migration. This innovative technology isstill in the early stages of development.

The EPA report Chemical Enhancements toPump-and-Treat Remediation (Palmer and Fish,1992) provides additional information on techni-cal issues related to this topic.

7.4. Biological EnhancementsBiological enhancements to pump-and-treat

systems stimulate subsurface microorganisms,primarily bacteria, to degrade contaminants toharmless mineral end products, such as carbondioxide and water. In situ bioremediation ofcertain types of hydrocarbons (primarily petro-leum products and derivatives), encouraged byaddition of oxygen and nutrients to the groundwater, is an established technology. Other readilybiodegradable substances, such as phenol, cresols,acetone, and cellulosic wastes, are also amenableto aerobic in situ bioremediation. Key elements insuch a system are delivery of oxygen and nutri-ents by use of an injection well (Figure 38a) or aninfiltration gallery (Figure 38b). A limitation of insitu bioremediation is that minimum contaminantconcentrations required to maintain microbialpopulations may exceed health-based cleanupstandards, particularly where heavier hydrocar-bons are involved.

In situ bioremediation of chlorinated solvents isless well demonstrated because metabolic pro-cesses for their degradation are more complexthan those for hydrocarbon degradation (NRC,1994). Nonetheless, methanotrophs are able todegrade some chlorinated solvents under aerobicconditions if methane is supplied as an energysource. Also, the ability of anaerobic bacteria todegrade a variety of chlorinated solvents is well

60

Figure 36.Process diagramfor air spargingwith (a) verticalwells, and (b)horizontal wells(after NRC,1994).

Vent toAtmosphere

Air Compressor

Vapor ExtractionWells

Vacuum

Direction of Ground-Water FlowStreamtubes of Air

VaporTreatment

Contaminated Zone

Surface Soil/Cap

Unsaturated Zone

Saturated Zone

Injection Point for Flushing Gas

Extraction of Contaminated Gas

Contaminated Zone

Surface Soil/Cap

Unsaturated Zone

Saturated Zone

(a)

(b)

61

Figure 37 .Schematic of chemicalenhancement of a pump-and-treat system. Key areas ofconcern are shown in boxes.In some cases, the reactiveagent will be recovered andreused (Palmer and Fish,1992).

documented. Two major obstacles to the use ofanaerobic processes for in situ bioremediation arethat (1) hazardous intermediate degradationproducts can accumulate, and (2) undesirablewater quality changes, such as dissolution of ironand manganese, can occur.

EPA reference sources identified at the end ofthis guide (Section 9) that are particularly relevantto in situ bioremediation include Norris et al.(1993), Sims et al. (1992), and U.S. EPA (1993,1994).

7.5. Alternatives to the Pump-and-Treat Approach

Nearly all approaches to ground-water cleanupinvolve some degree of ground-water pumping.Even when containment is the primary objective,low-flow pump-and-treat systems are usuallyrequired to prevent the escape of contaminatedwater from the confined area. Two remediationapproaches that eliminate pumping as a compo-nent of the system are (1) intrinsic bioremedia-tion, and (2) in situ reactive barriers. Althoughboth of these methods show promise, they are still

Delivery Reaction Removal

Injection of ReactiveAgents

Extraction ofReactive Agentsand Contaminants

Recovery ofReactive Agent

Treatment

DISPOSAL

62

Figure 38.Two types ofaerobic in situbioremediationsystems: (a)injection well withsparger, (b)infiltration gallery(Sims et al.,1992, afterThomas andWard, 1989).

Spilled Materials

Clay

Sparger

Injection Well

WaterSupply

Nutrient Addition TankTo Sewer orRecirculate

Air Compressor

(a)

Production Well

Water Table

Course Sand

Recirculated Waterand Nutrients

Recovery Well

Monitoring Well

WaterTable

Trapped Hydrocarbons

Infiltration Gallery

Air Compressor orHydrogen PeroxideTank Nutrient Addition

(b)

63

in development and their effectiveness remains tobe demonstrated.

7.5.1. Intrinsic BioremediationIntrinsic bioremediation relies on indigenous

microbes to biodegrade organic contaminants,without human intervention in the form of supply-ing electron acceptors, nutrients, and othermaterials. The processes that occur are the sameas those in engineered bioremediation systems,but they occur more slowly. A decision to refrainfrom active site manipulation does not eliminatethe need to conduct ground-water sampling withinthe contaminant plume to document that biodeg-radation is occurring. Moreover, sampling wouldstill need to be performed outside the contami-nated area to identify any offsite migration ofcontaminants that might require initiation of moreactive remedial measures (Figure 39b). There is agreater risk of failure with intrinsic bioremedia-tion compared to engineered bioremediationbecause no active measures are used to control thecontaminant plume. The possible perception thatintrinsic bioremediation is the equivalent to doingnothing is also a barrier to its acceptance (NRC,1994).

7.5.2. In Situ Reactive BarriersThe concept of using permeable in situ reactive

barriers to treat a contaminant plume as it movesthrough an aquifer under natural hydraulicgradients (Figure 39c and 39d) was first suggestedby McMurty and Elton (1985), but it has onlyrecently begun to receive significant attentionfrom the research community (Starr and Cherry,1994). The funnel-and-gate concept, whichcombines impermeable barriers to contain andchannel the flow of the contaminant plume towardthe reactive barrier has received the most attentionbecause numerous possible configurations can bedeveloped to address different types of contami-nant plumes and geologic settings (Figure 40).Depending on the contaminants present in theplume, the reactive zone uses a combination ofphysical, chemical, and biological processes.

The great promise of in situ reactive barriers isthat they will require little or no energy input onceinstalled, yet provide more active control andtreatment of the contaminant plume than intrinsicbioremediation. The main engineering challengesinvolve provision of suitable amounts of reactivematerials in a permeable medium and properplacement to avoid short-circuiting the contactbetween the gate and the cutoff wall.

64

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

Extraction Well

Cutoff Wall'Funnel'

In Situ Reactor'Gate'

In Situ Reaction Curtain'

MonitoringWells

Ground-Water Plume

Contaminant SourceZone

Remediated Plume

(a)

(b)

(c)

(d)

Remediated Plume

65

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

Single Gate System

Multiple Gate System Multiple Reactor Systems

Fully Penetrating Gate Hanging Gate

(a)

(b)

(c)

66

8. ReferencesCohen, R.M., A.H. Vincent, J.W. Mercer, C.R.

Faust, and C.P. Spalding. 1994. Methods forMonitoring Pump-and-Treat Performance.EPA/600/R-94/123. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 102 pp.

Clausen, J.L. and D.A. Solomon. 1994. Character-ization of Ground Water Plumes and DNAPLSources Using a Driven Discreet-DepthSampling System. Ground Water Manage-ment 18:435-445 (Proc. of 8th Nat. OutdoorAction Conf. on Aquifer Remediation,Ground Water Monitoring and GeophysicalMethods).

Feldman, P.R. and D.J. Campbell. 1994. Evaluat-ing the Technical Impracticality of Ground-Water Cleanup. Ground Water Management18:595-608 (Proc. of 8th Nat. Outdoor ActionConf. on Aquifer Remediation, Ground WaterMonitoring and Geophysical Methods).

Freeze, R.A. and J.A. Cherry. 1989. What HasGone Wrong? Ground Water 27(4):458-464.

Gillham, R.W., E.A. Sudicky, J.A. Cherry, andE.O. Frind. 1984. An Advective-DiffusionConcept to Solute Transport in HeterogeneousUnconsolidated Geologic Deposits. WaterResour. Res. 20(3):369-378.

Gorelick, S.M., R.A. Freeze, D. Donohue, andJ.F. Keely. 1993. Groundwater Contamina-tion: Optimal Capture and Containment.Lewis Publishers: Boca Raton, FL. 416 pp.

Bartow, G. and C. Davenport. 1995. Pump-and-Treat Accomplishments: A Review of theEffectiveness of Ground Water Remediationin Santa Clara Valley, California. GroundWater Monitoring and Remediation15(2):140-146.

Berglund, S. and V. Cvetkovic. 1995. Pump-and-Treat Remediation of Heterogeneous Aqui-fers: Effects of Rate-Limited Mass Transfer.Ground Water 33(4):675-685.

Blandford, T.N. and P.S. Huyakorn. 1991. WHPA:Modular Semi-Analytical Model for theDelineation of Wellhead Protection Areas,Version 2.0. Office of Ground Water Protec-tion; Available from EPA Center for Subsur-face Modeling Support, Ada, OK. Version 1.0was released in 1990 [Four modules:MWCAP, RESSQC, GPTRAC, MONTEC;most current disk version is 2.1]

Bradbury, K.R., M.A. Muldoon, A. Zaporozec,and J. Levy. 1991. Delineation of WellheadProtection Areas in Fractured Rocks. EPA/570/9-91-009. Office of Water, Washington,DC. 144 pp.

Cohen, R.M. and J.W. Mercer. 1993. DNAPL SiteEvaluation. EPA/600/R-93/002 (NTIS PB93-150217). R.S. Kerr Environmental ResearchLaboratory, Ada, OK. [Also published byLewis Publishers as C.K. Smoley edition,Boca Raton, FL. 384 pp.]

67

Marquis, Jr., S. 1995. Don’t Give Up on Pumpand Treat: Enhance It with Bioremediation.Soils & Groundwater Cleanup, August-September, pp. 46-50.

McMurty, D.C., and R.O. Elton. 1985. NewApproach to In-Situ Treatment of Contami-nated Groundwaters. Environ. Progress4(3):168-170.

National Research Council (NRC). 1994. Alterna-tives for Ground Water Cleanup. NationalAcademy Press. 336 pp.

Norris, R.D. et al. 1993. In-Situ Bioremediationof Ground Water and Geological Material: AReview of Technologies. EPA/600/R-93/124(NTIS PB93-215564). R.S. Kerr Environmen-tal Research Laboratory, Ada, OK. [13authors; see also Norris et al., 1994]

Palmer, C.D. and W. Fish. 1992. ChemicalEnhancements to Pump-and-Treat Remedia-tion. Ground Water Issue Paper. EPA/540/S-92/001. R.S. Kerr Environmental ResearchLaboratory, Ada, OK. 20 pp.

Papadopulos & Associates, Inc. and Conestoga-Rovers & Associates Ltd. 1993. Chem-DyneSite Trust Fund; 1992 Annual Report. Chem-Dyne Site, Hamilton, OH. April.

Piwoni, M.D. and J.W. Keeley. 1990. BasicConcepts of Contaminant Sorption at Hazard-ous Waste Sites. Ground Water Issue. EPA/540/4-90/053. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 7 pp.

Rogers, L.L., R.U. Dowla, and V.M. Johnson.1995. Optimal Field-Scale GroundwaterRemediation Using Neural Networks andGenetic Algorithm. Environ. Sci. Technol.29(5):1145-1155.

Haitjema, H.M., J. Wittman, V. Kelson, and N.Bauch. 1994. WhAEM: Program Documenta-tion for the Wellhead Analytic ElementModel. EPA/600/R-94/210, 120 pp. Availablefrom EPA Center for Subsurface ModelingSupport, Ada, OK. [Includes GeographicAnalytic Element Preprocessor (GAEP) andCapture Zone Analytic Element Model(CZAEM)]

Haley, J.L., B. Hanson, C. Enfield, and J. Glass.1991. Evaluating the Effectiveness of GroundWater Extraction Systems. Ground WaterMonitoring Rev. 11(1):119-124. [Summary ofU.S. EPA (1989)]

Harvey, C.F., R. Haggerty, and S.M. Gorelick.1994. Aquifer Remediation: A Method forEstimating Mass Transfer Rate Coefficientsand an Evaluation of Pulsed Pumping. WaterResour. Res. 30(7):1979-1991.

Hoffman, F. 1993. Ground-Water RemediationUsing “Smart Pump and Treat.” GroundWater 31(1):98-106.

Keely, J.F. 1989. Performance Evaluation ofPump-and-Treat Remediations. SuperfundIssue Paper. EPA/540/8-89/005. R.S. KerrEnvironmental Research Laboratory, Ada,OK. 14 pp.

Knox, R.C., L.W. Canter, D.F. Kincannon, E.L.Stover, and C.H. Ward. 1984. State-of-the Artof Aquifer Restoration. EPA/600/2-84/182a&b (National Technical InformationService [NTIS] PB85-181071 and PB85-181089). R.S. Kerr Environmental ResearchLaboratory, Ada, OK.

Mackay, D.M. and J.A. Cherry. 1989. Groundwa-ter Contamination: Pump-and-Treat Remedia-tion. Environ. Sci. Technol. 23(6):630-636.

68

Roy, W.R., I.G. Krapac, S.F.J. Chou, and R.A.Griffin. 1992. Batch-Type Procedures forEstimating Soil Adsorption of Chemicals.EPA/530/SW-87/006F (NTIS PB92-146190).Risk Reduction Engineering Laboratory,Cincinnati, OH. 100 pp.

Satkin, R.L. and P.B. Bedient. 1988. Effectivenessof Various Aquifer Restoration SchemesUnder Variable Hydrogeologic Conditions.Ground Water 26(4):488-498.

Sims, J.L., J.M. Suflita, and H.H. Russell. 1992.In-Situ Bioremediation of ContaminatedGround Water. Ground Water Issue Paper.EPA/540/S-92/003. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 11 pp.

Starr, R.C. and J.A. Cherry. 1994. In Situ Reme-diation of Contaminated Ground Water: TheFunnel-and-Gate System. Ground Water32(3):465-476.

Strack, O.D.L. et al. 1994. CZAEM User’s Guide:Modeling Capture Zones of Ground-WaterWells Using Analytic Elements. EPA/600/R-94/174, 58 pp. Available from EPA Center forSubsurface Modeling Support, Ada, OK. [Seealso, Haitjema et al., 1994]

Thomas, J.D., and C.H. Ward. 1989. In SituBiorestoration of Organic Contaminants inthe Subsurface. Environ. Sci. Technol.23:760-786.

Travis, C.C. and C.B. Doty. 1990. Can Contami-nated Aquifers at Superfund Site BeRemediated? Environ. Sci. Technol.24(1):1464-1466.

U.S. Environmental Protection Agency (EPA).1988. Guidance on Remedial Actions forContaminated Ground Water at Superfund

Sites. EPA/540/G-88/003. Office of SolidWaste and Emergency Response (OSWER)Directive 9283.1-2 (NTIS PB89-184618).Office of Solid Waste and Emergency Re-sponse, Washington, DC.

U.S. Environmental Protection Agency (EPA).1989. Evaluation of Ground-Water ExtractionRemedies: Volume 1, Summary Report (EPA/540/2-89/054, NTIS PB90-183583, 66 pp.);Volume 2, Case Studies 1-19 (EPA/540/2-89/054b); and Volume 3, General Site Data BaseReports (EPA/540/2-89/054c). Office of SolidWaste and Emergency Response, Washington,DC.

U.S. Environmental Protection Agency (EPA).1991. Handbook: Stabilization Technologiesfor RCRA Corrective Actions. EPA/625/6-91/026. Center for Environmental ResearchInformation, Cincinnati, OH. 62 pp.

U.S. Environmental Protection Agency (EPA).1992. Methods for Evaluating the Attainmentof Cleanup Standards, Volume 2: GroundWater. EPA/230/R-92/014. Office of SolidWaste and Emergency Response, Washington,DC.

U.S. Environmental Protection Agency (EPA).1993. Guidance for Evaluation the TechnicalImpracticability of Ground-Water Restora-tion. EPA/540/R-93/080, OSWER 0234.2-25(NTIS PB93-963507). Office of Solid Wasteand Emergency Response, Washington, DC.

U.S. Environmental Protection Agency (EPA).1994. Manual: Alternative Methods for FluidDeliver and Recovery. EPA/625/R-94/003.Center for Environmental Research Informa-tion, Cincinnati, OH. 87 pp.

69

U.S. Environmental Protection Agency (EPA).1995. Manual: Ground-Water and LeachateTreatment Systems. EPA/625/R-94/005.Center for Environmental Research Informa-tion, Cincinnati, OH. 119 pp.

Wilson, J.L., S.H. Conrad, W.R. Mason, W.Peplinski and E. Hagen. 1990. LaboratoryInvestigation of Residual Liquid Organics.EPA/600/6-90/004. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 267 pp.

70

9. EPA Publications Providing Further InformationThe EPA publications listed below provide more

detailed information on the subjects discussedin this document. Publications and additionalcopies of this brochure can be obtained at nocharge (while supplies are available) from thefollowing sources:

EPA/625-series documents: Office of Researchand Development (ORD) Publications, P.O.Box 19968, Cincinnati, OH 45219-0968;phone 513 569-7562, fax 513 569-7562.

Other EPA documents: National Center forEnvironmental Publications and Information(NCEPI), 11029 Kenwood Road, Cincinnati,OH 45242; fax 513 891-6685.

Other documents, for which an NTIS acquisitionnumber is shown can be obtained from theNational Technical Information Service(NTIS), Springfield, VA 22161; 800 336-4700, fax 703/321-8547.

Contaminant Transport and FateHuling, S.G. 1989. Facilitated Transport. Ground

Water Issue. EPA/540/4-89/003. R.S. KerrEnvironmental Research Laboratory, Ada,OK. 5 pp.

Huling, S.C. and J.W. Weaver. 1991. DenseNonaqueous Phase Liquids. Ground WaterIssue. EPA/540/4-91/002. R.S. Kerr Environ-mental Research Laboratory, Ada, OK. 21 pp.

McLean, J.E. and B.E. Bledsoe. 1992. Behaviorof Metals in Soils. Ground Water Issue. EPA/540/S-92/018. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 25 pp.

Palmer, C.D. and R.W. Puls. 1994. NaturalAttenuation of Hexavalent Chromium inGround Water and Soils. Ground Water Issue.EPA/540/S-94/505. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 13 pp.

Piwoni, M.D. and J.W. Keeley. 1990. BasicConcepts of Contaminant Sorption at Hazard-ous Waste Sites. Ground Water Issue. EPA/540/4-90/053. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 7 pp.

Sims, J.L., J.M. Suflita, and H.H. Russell. 1991.Reductive Dehalogenation of Organic Con-taminates in Soils and Ground Water. GroundWater Issue. EPA/540/4-91/054. R.S. KerrEnvironmental Research Laboratory, Ada,OK. 12 pp.

Wilson, J.L., S.H. Conrad, W.R. Mason, W.Peplinski, and E. Hagen. 1990. LaboratoryInvestigation of Residual Liquid Organics.EPA/600/6-90/004. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 267 pp.

Site CharacterizationU.S. Environmental Protection Agency (EPA).

1991. Site Characterization for SubsurfaceRemediation. EPA/625/4-91/026. Center for

71

Environmental Research Information, Cincin-nati, OH. 259 pp.

U.S. Environmental Protection Agency (EPA).1992. Estimating the Potential for the Occur-rence of DNAPL at Superfund Sites. OSWERPublication 9355.4-07/FS. Office of SolidWaste and Emergency Response, Washington,DC.

U.S. Environmental Protection Agency (EPA).1993. Evaluation of the Likelihood ofDNAPL Presence at NPL Sites. EPA/540/R-93/002 (OSWER 0355.4-13). Office of SolidWaste and Emergency Response, Washington,DC.

U.S. Environmental Protection Agency (EPA).1993. Use of Airborne, Surface and BoreholeGeophysical Techniques at ContaminatedSites: A Reference Guide. EPA/625/R-92/007.Center for Environmental Research Informa-tion, Cincinnati, OH.

U.S. Environmental Protection Agency (EPA).1993. Subsurface Characterization andMonitoring Techniques: A Desk ReferenceGuide; Vol. I: Solids and Ground Water; Vol.II: The Vadose Zone, Field Screening andAnalytical Methods. EPA/625/R-93/003a&b.Center for Environmental Research Informa-tion, Cincinnati, OH.

Pump-and-Treat SystemsCohen, R.M., A.H. Vincent, J.W. Mercer, C.R.

Faust, and C.P. Spalding. 1994. Methods forMonitoring Pump-and-Treat Performance.EPA/600/R-94/123. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 102 pp.

Keely, J.F. 1989. Performance Evaluation ofPump-and-Treat Remediations. SuperfundIssue Paper. EPA 540/8-89/005. R.S. Kerr

Environmental Research Laboratory, Ada,OK. 14 pp.

Mercer, J.W., D.C. Skipp, and D. Giffin. 1990.Basics of Pump-and-Treat Ground-WaterRemediation Technology. EPA/600/8-90/003.R.S. Kerr Environmental Research Labora-tory, Ada, OK. 58 pp.

Palmer, C.D. and W. Fish. 1992. ChemicalEnhancements to Pump-and-Treat Remedia-tion. Ground Water Issue Paper. EPA/540/S-92/001. R.S. Kerr Environmental ResearchLaboratory, Ada, OK. 20 pp.

Repa, E. and D.P. Doerr. 1985. Leachate PlumeManagement. EPA/540/2-85/004 (NTISPB86-122330). Hazardous Waste EngineeringResearch Laboratory, Cincinnati, OH.

U.S. Environmental Protection Agency (EPA).1989. Evaluation of Ground-Water ExtractionRemedies: Volume 1, Summary Report (EPA/540/2-89/054, NTIS PB90-183583, 66 pp.);Volume 2, Case Studies 1-19 (EPA/540/2-89/054b); and Volume 3, General Site Data BaseReports (EPA/540/2-89/054c). Office of SolidWaste and Emergency Response, Washington,DC.

U.S. Environmental Protection Agency (EPA).1992. Evaluation of Ground-Water ExtractionRemedies, Phase II. Oswer Publication9355.4-05, Vols. 1-2. Office of Solid Wasteand Emergency Response, Washington, DC.

U.S. Environmental Protection Agency (EPA).1992. General Methods for Remedial Opera-tions Performance Evaluation. EPA/600/R-92/002. R.S. Kerr Environmental ResearchLaboratory, Ada, OK. 37 pp.

U.S. Environmental Protection Agency (EPA).1993. Guidance for Evaluating the Technical

72

Impracticability of Ground-Water Restora-tion. EPA/540/R-93/080, OSWER 0234.2-25(NTIS PB93-963507). Office of Solid Wasteand Emergency Response, Washington, DC.

U.S. Environmental Protection Agency (EPA).1993. Bioremediation Resource Guide. EPA/542/B-93/004. Office of Solid Waste andEmergency Response, Washington, DC.[Includes annotated list of more than 80significant references]

U.S. Environmental Protection Agency (EPA).1994. Bioremediation in the Field. EPA/540/N-94/501. Office of Solid Waste and Emer-gency Response, Washington, DC. [Periodi-cally updated; latest issue No. 11, July, 1994]

U.S. Environmental Protection Agency (EPA).1994. Manual: Alternative Methods for FluidDelivery and Recovery. EPA/625/R-94/003.Center for Environmental Research Informa-tion, Cincinnati, OH. 87 pp.

U.S. Environmental Protection Agency (EPA).1995. In Situ Remediation Technology StatusReport: Hydraulic and Pneumatic Fracturing.EPA/542/K-94/005. Office of Solid Waste andEmergency Response, Washington, DC. 15pp.

Ground-Water Treatment MethodsCanter, L.W. and R.C. Knox. 1986. Ground Water

Pollution Control. Lewis Publishers: Chelsea,MI. 526 pp. [Contains mostly same materialas Knox et al. (1984)]

Knox, R.C., L.W. Canter, D.F. Kincannon, E.L.Stover, and C.H. Ward. 1984. State-of-the Artof Aquifer Restoration. EPA 600/2-84/182a&b (NTIS PB85-181071 and PB85-181089). R.S. Kerr Environmental ResearchLaboratory, Ada, OK. [See also Canter andKnox (1985)]

McArdle, J.L., M.M. Arozarena, and W.E.Gallagher. 1987. A Handbook on Treatmentof Hazardous Waste Leachate. EPA/600/8-87/006 (NTIS PB87-152328). Hazardous WasteEngineering Research Laboratory, Cincinnati,OH.

U.S. Department of Defense EnvironmentalTechnology Transfer Committee (DOD/ETTC). 1994. Remediation TechnologiesScreening Matrix and Reference Guide. EPA/542/B-94/013 (NTIS PB95-104782). Officeof Solid Waste and Emergency Response,Washington, DC.

U.S. Environmental Protection Agency (EPA).1994. Ground-Water Treatment TechnologyResource Guide. EPA/542/B-94/009. Officeof Solid Waste and Emergency Response,Washington, DC. [Includes annotated list ofmore than 60 significant references]

U.S. Environmental Protection Agency (EPA).1994. Innovative Treatment TechnologiesAnnual Status Report, 6th ed. EPA/542/R-94/005. Office of Solid Waste and EmergencyResponse, Washington, DC.

U.S. Environmental Protection Agency (EPA).1994. Superfund Innovative TechnologyEvaluation Program: Technology Profiles, 7thed. EPA/540/R-94/526. Risk ReductionEngineering Laboratory, Cincinnati, OH. 499pp.

U.S. Environmental Protection Agency (EPA).1995. Manual: Ground-Water and LeachateTreatment Systems. EPA/625/R-94/005.Center for Environmental Research Informa-tion, Cincinnati, OH.

73

In Situ Ground-Water TreatmentNorris, R.D. et al. 1993. In-Situ Bioremediation

of Ground Water and Geological Material: AReview of Technologies. EPA/600/R-93/124(NTIS PB93-215564). R.S. Kerr Environmen-tal Research Laboratory, Ada, OK.[13 au-thors; see also Norris et al., 1994]

Norris, R.D. et al. 1994. Handbook of Bioreme-diation. Boca Raton, FL: Lewis Publishers.272 pp. [Contains same material as Norris etal., 1993]

Sims, J.L., J.M. Suflita, and H.H. Russell. 1992.In Situ Bioremediation of ContaminatedGround Water. Ground Water Issue Paper.EPA/540/S-92/003. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 11 pp.

U.S. Environmental Protection Agency (EPA).1995a. In Situ Remediation TechnologyStatus Report: Thermal Enhancements. EPA/542/K-94/009. Office of Solid Waste andEmergency Response, Washington, DC. 22pp.

U.S. Environmental Protection Agency (EPA).1995b. In Situ Remediation TechnologyStatus Report: Surfactant Enhancements.EPA/542/K-94/003. Office of Solid Waste andEmergency Response, Washington, DC. 22pp.

U.S. Environmental Protection Agency (EPA).1995c. In Situ Remediation TechnologyStatus Report: Cosolvents. EPA/542/K-94/006. Office of Solid Waste and EmergencyResponse, Washington, DC. 6 pp.

U.S. Environmental Protection Agency (EPA).1995d. In Situ Remediation TechnologyStatus Report: Electrokinetics. EPA/542/K-94/007. Office of Solid Waste and EmergencyResponse, Washington, DC. 20 pp.

U.S. Environmental Protection Agency (EPA).1995e. In Situ Remediation TechnologyStatus Report: Treatment Walls. EPA/542/K-94/004. Office of Solid Waste and EmergencyResponse, Washington, DC. 26 pp.

Ground-Water ModelingBear, J., M.S. Beljin, and R.R. Ross. 1992.

Fundamentals of Ground-Water Modeling.Ground Water Issue. EPA/540/S-92/005. R.S.Kerr Environmental Research Laboratory,Ada, OK. 11 pp.

Schmelling, S.G. and R.R. Ross. 1989. Contami-nant Transport in Fractured Media: Modelsfor Decisionmakers. Ground Water Issue.EPA/540/4-89/004. R.S. Kerr EnvironmentalResearch Laboratory, Ada, OK. 8 pp.

U.S. Environmental Protection Agency (EPA).1988. Selection Criteria for MathematicalModels Used in Exposure Assessments:Ground-Water Models. EPA/600/8-88/075(NTIS PB88-248752). Office of Health andEnvironmental Assessment, Washington, DC.[Contains summary tables and descriptions of63 analytical solutions and 49 analytical andnumerical codes for evaluating ground-watercontaminant transport]

U.S. Environmental Protection Agency (EPA).1994. Assessment Framework for Ground-Water Model Applications. EPA/500/B-94/003 (OSWER Directive 9029.00). Office ofSolid Waste and Emergency Response,Washington, DC. 41 pp.

van der Heijde, P.K.M. 1994. Identification andCompilation of Unsaturated/Vadose ZoneModels. EPA/600/R-94/028 (NTIS PB94-157773). R.S. Kerr Environmental ResearchLaboratory, Ada, OK.

74

van der Heijde, P.K.M. and O.A. Einawawy. 1993.Compilation of Ground-Water Models. EPA/600/R-93/118 (NTIS PB93-209401). R.S.Kerr Environmental Research Laboratory,

Ada, OK. [Summary information on modelsfor porous media flow and transport,hydrogeochemical models, stochastic models,and fractured rock]