vvEPA
United States
Environmental Protection
Agency
Office of Emergency and
Remedial Response
Washington, DC 20460
EPA/540.2-89/054
September 1989
Superfund
Evaluation of Ground-
Water Extraction
Remedies
Volume 1
Summary Report
-------
EPA/540/2-89/054
September 1989
Evaluation of Ground-Water
Extraction Remedies
Volume 1
Summary Report
Agency
. .
' • -.on Boulevard, 12th Floor
. . 60604-3590 --
Office of Emergency and Remedial Response
U.S. Environmental Protection Agency
Washington, DC 20460
-------
Notice
Development of this document was funded by the United States Environmental
Protection Agency in part under contract No. 68-W8-0098 to CH2M HILL
SOUTHEAST. It has been sub|ected to the Agency's review process and approved for
publication as an EPA document.
The policies and procedures set out in this document are intended solely for the
guidance of response personnel. They are not intended, nor can they be rened uoon.
to create any ngnts, substantive or procedural, enforceable by any party in litigation
with the United States. The Agency reserves the right to act at variance witn these
policies and procedures and to change them at any time without public notice.
-------
CONTENTS
Page
Executive Summary E-l
1 Introduction 1-1
1.1 Purpose of the Study 1-1
1.2 Study Methods 1-2
1.2.1 Collection of General Data 1-3
1.2.2 Collection of Detailed Site Data 1-3
1.3 Limitations of Data Collection 1-4
2 Overview of Results 2-1
2.1 General Site Data Collection 2-1
2.1.1 Site Locations 2-1
2.1.2 Administrative Programs 2-2
2.1.3 Remediation Objectives 2-2
2.1.4 Type of Extraction System 2-3
2.1.5 Extraction System Enhancements 2-3
2.1.6 Type of Site 2-3
2.1.7 Type of Contaminant 2-4
2.1.8 Presence of Nonaqueous Liquids 2-4
2.1.9 Geologic Materials 2-5
2.1.10 Implementation Status 2-5
2.2 Detailed Data Collection and Case Studies 2-5
2.2.1 Remediation Objectives 2-6
2.2.2 Contaminant Characteristics 2-7
2.2.3 Geologic Environments 2-8
2.2.4 Use of Innovative Technologies 2-9
2.2.5 Extraction System Design
Information 2-10
2.3 General Conclusions 2-11
2.3.1 Aquifer Restoration 2-14
2.3.2 Migration Control 2-15
2.3.3 Well-Head Treatment 2-16
-------
Page
3 Factors Affecting System Design and
Performance 3-1
3.1 Factors Affecting Migration Control 3-1
3.1.1 Contaminant Distribution 3-1
3.1.2 Effects of Pre-Existing
Gradients 3-2
3.1.3 Effects of Aquifer Properties 3-3
3.1.3.1 Transmissivity 3-3
3.1.3.2 Heterogeneity 3-3
3.2 Factors Affecting Aquifer Restoration 3-4
3.2.1 Well Placement and Pumping Rate 3-5
3.2.2 Contaminant Sorption and
Retardation 3-6
3.2.3 Isolation of Low Permeability
Zones 3-8
3.2.4 Contaminants in Nonaqueous Form
(NAPLs) 3-8
3.2.5 Leaching of Contaminants
from the Vadose Zone 3-9
4 Information Requirements for System Design and
Operation 4-1
4.1 Hydrogeologic Information 4-1
4.1.1 Stratigraphy 4-1
4.1.2 Aquifer Hydraulic Properties 4-2
4.1.3 Potentiometric Gradients 4-3
4.2 Contaminant Distribution and
Characteristics 4-4
4.2.1 Identification of Contaminants 4-4
4.2.2 Contaminant Distribution and
Concentration 4-5
4.2.3 Contaminant Mobility
Characteristics 4-7
4.2.4 Identification of Contaminant
Sources 4-7
-------
Page
4.3 Performance Monitoring Requirements 4-8
4.4 Post-Termination Monitoring 4-9
5. Methods of Increasing System Effectiveness 5-1
5.1 Progressive System Modification 5-1
5.2 Pulsed Pumping 5-2
5.3 Physical Containment Systems 5-2
5.4 Reinjection 5-3
5.5 Vapor Extraction 5-3
5.6 Fracture Enhancement 5-4
Appendix
A Properties of Some Common Contaminants
B Contaminant Partitioning in Sorbing Media
VOLUME 2—Case Studies (Bound Separately)
VOLUME 3--General Site Data—Data Base Reports (Bound
Separately)
WDCR161/015.50
-------
EXECUTIVE SUMMARY
Ground-water extraction is the most commonly used remedial
technology for contaminated aquifers. In this
investigation, information is assembled from hazardous waste
sites throughout the United States showing how ground-water
extraction systems are being used, how their performance
compares with expectations, and what factors are affecting
their success.
STUDY METHODS
The study is based on an information gathering effort
consisting of two components. In the first component,
general information was collected about hazardous waste
sites where ground-water extraction is either planned or is
already in use. This general data was obtained from EPA
regional offices, state agencies, and environmental
consultants. General data describing the site locations,
the types of contaminants involved, the geologic nature of
the sites, and the status of the remediations were collected
for a total of 112 sites. Most of these sites were still in
the pre-implementation phase or had generated such a short
record of extraction system performance that no evaluation
could be made. The 112 data base reports for these sites
are bound separately as Volume 3 of this report.
The second component of the data gathering effort was to
obtain detailed information on selected sites where ground-
water extraction systems are already in operation. The
information required included detailed descriptions of site
hydrogeology and contaminant distributions, the design of
the extraction systems and the analyses used in design, and
data on the performance of the systems. Detailed
information was obtained for 18 sites, and was used to
prepare a series of 18 case studies, which are bound
separately as Volume 2 of this study. Also included in
Volume 2 is a 19th case study for a site in Canada that was
prepared by the Environmental Ministry of Quebec.
STUDY FINDINGS
Trends identified from the 19 case studies lead to the
following general conclusions :
o The ground-water extraction systems were generally
effective in maintaining hydraulic containment of
contaminant plumes, thus preventing further
migration of contaminants.
o Significant removal of contaminant mass from the
subsurface is often achieved by ground-water
extraction systems. When site conditions are
favorable and the extraction system is properly
The policy implications of these findings are discussed in
a separate memorandum from the EPA's Acting Assistant
Administrator of the Office of Solid Waste and Emergency Response
(OSWER Directive 9355.4-03).
E-l
-------
extraction systems. When site conditions are
favorable and the extraction system is properly
designed and operated, it may be possible to
remediate the aquifer to health-based levels.
o Contaminant concentrations usually decrease most
rapidly soon after the initiation of extraction.
After this initial reduction, the concentrations
often tend to level off and progress toward
complete aquifer restoration is usually slower
than expected.
o Data collection, both prior to system design and
during operation, was frequently not sufficient to
fully assess contaminant movement and the response
of the ground-water system to extraction.
Three different remedial objectives have been identified for
the ground-water extraction systems described in the case
studies: aquifer remediation, migration control, and well-
head treatment. Aquifer restoration for the purposes of
this report means that the contaminant concentrations in the
aquifer are to be reduced below specified levels that have
been determined to be protective for the site. In the case
of Superfund sites, the cleanup levels are either the
regulatory Maximum Contaminant Levels (MCLs) or ICf4 to 10~6
excess cancer risk concentrations. Of the 19 sites studied
in detail, 13 had aquifer restoration as their primary goal,
and only 1 has been successful so far. Several of the other
systems show promise of eventual aquifer restoration, but
typically progress toward this goal is behind schedule.
Concentrations often decline rapidly when the extraction
system is first turned on, but after the initial decrease
continued reductions are usually slower than expected.
The operational experience described in the case studies
indicates that success in aquifer restoration depends on
favorable aquifer and contaminant characteristics, and on
appropriate system design. Sites that are favorable for
aquifer restoration have relatively simple stratigraphy with
fairly homogeneous unconsolidated aquifer materials and
contaminants that are present primarily as dissolved
constituents in the ground water. Most departures from
these ideal conditions tend to impede the progress of
aquifer restoration. However, even if the concentrations
are not rapidly reduced to cleanup goals, the extraction
systems may still significantly reduce contaminant mass in
the aquifer.
Migration control, or plume containment, was the primary
goal at 7 of the 19 case study sites. In some of these
cases the systems were initially intended for aquifer
restoration, but operating experience indicated that this
goal was not feasible. Migration control is a less
ambitious objective, but one that can be more easily
attained. In all but three of the 19 case study sites,
successful plume containment has been demonstrated. Failure
to contain the contaminant plume has been definitely
E-2
-------
Two of the sites described in the case studies involve well-
head treatment. These are sites where water-supply wells
have become contaminated, but have not been taken out of
service. Instead, treatment systems have been installed at
the wells to make the ground water produced suitable for its
intended use. Both of the systems studied are successfully
delivering water for domestic supply, and it appears that
well-head treatment can usually be operated successfully.
In addition to serving the primary need for water supply,
these systems can often help to prevent the spread of
contamination in the aquifer beyond the region already
affected.
The design of an extraction system should be based on
thorough analysis of the patterns of ground-water flow and
contaminant transport that the system will produce.
Frequently, the site information generated in pre-design
investigations is inadequate to support the analysis
necessary for optimal design. At a few of the sites
studied, the administration of the remedial program was
flexible enough to permit modification of the system in
response to site information produced by performance
monitoring.
The case studies also showed that monitoring of ground-water
extraction system performance is frequently inadequate.
Hydrologic monitoring is necessary for both aquifer
restoration and migration control systems to ensure that the
intended hydraulic capture zone is being maintained.
Contaminant concentrations must be monitored by taking
samples regularly from dedicated monitoring wells. Sampling
the ground water produced from the recovery wells does not
provide the information needed to evaluate the performance
of ground-water extraction systems.
WDCR161/014.50
E-3
-------
Chapter 1
INTRODUCTION
1.1 PURPOSE OF THE STUDY
Ground-water extraction is the most commonly used remedial
technology for contaminated aquifers. This method of
remediation, also referred to as the "pump and treat
method," usually includes three steps: (1) extraction of
contaminated ground water from the aquifer via recovery
wells, (2) treatment of the extracted water, and
(3) disposal of the contaminants and discharge of the
treated water. This study is concerned primarily with the
first step in this process, the extraction of the
contaminated ground water.
The specific goals of the study are:
1. To assemble information on how and where ground-water
extraction is being used
2. To evaluate the performance of extraction systems at
sites where operational experience is available
3. To investigate the design and operational
considerations that have affected the performance of
the extraction systems at these sites
As more experience has been gained with the long-term
operation of ground-water extraction systems, it has become
apparent that their performance often does not meet initial
expectations. This is particularly true of systems that
have been installed with the intention of cleaning up
contaminated aquifers to health-based concentration goals.
Cases where performance goals for aquifer cleanup have been
met or exceeded are quite rare. As a result, questions have
been raised concerning the general feasibility of aquifer
remediation to health-based standards, how it is affected by
site conditions, and how long remediation should be expected
to take. This study has identified a few sites where
ground-water extraction has either achieved the remediation
goals or appears to be approaching those goals on schedule.
More commonly, sites have been identified where progress
toward remediation is lagging behind expectations, or the
goal of remediation appears to be impractical and has been
replaced with a containment objective.
Several physical phenomena have been identified that tend to
interfere with the aquifer cleanup performance of extraction
1-1
-------
systems. One of them is adsorptive partitioning of
contaminants between the ground water and the aquifer
materials. If adsorption in neglected, projections of
extraction system effectiveness are likely to be over-
optimistic. Aquifer heterogeneity can also reduce
extraction system effectiveness by making it difficult to
control the ground-water flow patterns in the area to be
remediated. Aquifer cleanup where residual contaminant
sources are present is difficult and often infeasible.
Examples of residual contaminant sources include
unremediated disposal areas, contaminated soils above the
water table, and the presence of contaminants in a
nonaqueous phase.
Aquifer restoration is not the only objective of ground-
water extraction systems. Another commonly expressed goal
is migration control, or hydraulic containment of the
contaminant plume. In most aquifer restoration systems,
plume containment is listed as a secondary goal. Indeed, it
is usually necessary to establish control of contaminant
migration if the aquifer is to be cleaned up. (Exceptions
to this are sites where the aquifer can restore itself
naturally by discharging to surface water bodies or through
chemical or biological degradation of the ground-water
contaminants.)
Another form of ground-water extraction, called the well-
head treatment system, is sometimes found at well fields
where production wells have become contaminated. Here it
has been found to be cost-effective to continue producing
contaminated ground water but to remove the contaminants by
treatment before delivering it to the users. There are
several variations of this approach. At some sites the
source of the contamination is known, and an auxiliary
extraction system has been installed there. This auxiliary
system may be intended either to clean up the contaminated
aquifer or simply to prevent continued migration toward the
well field. In other cases, the source of the contamination
may be unknown, and the well-head treatment system may be
the only practical alternative.
1.2 STUDY METHODS
This is an empirical study based on the collection of
information from actual ground-water contamination sites.
It involves two separate but complementary data collection
components: collection of general data on the use of ground-
water extraction as a remedial technology, and detailed data
collection for selected sites with operating ground-water
1-2
-------
extraction systems. Information from the general data
collection effort was used to select sites that were
suitable for further detailed study.
1.2.1 Collection of General Data
The general data collection effort was intended to get a
limited amount of information on a large number of sites.
Its purpose was to provide general information on the
locations of ground-water extraction sites, their remedial
objectives, the hydrologic conditions and types of
contaminants involved, the administrative programs
controlling them, and the present status of the projects.
The candidate sites for general data collection were not
limited to those with operating systems, but also included
systems that are still in the design or installation phase.
However, information was not collected for sites
contaminated by leaking fuel tanks or fuel spills. These
sites were not included because they are so numerous and
because the special problems associated with floating
nonaqueous product deserve consideration at a level that is
beyond the scope of this investigation.
Information was collected from EPA staff members involved
with hazardous waste and RCRA sites as well as from state
environmental protection offices in California, Florida,
Minnesota, New Jersey, and New York. Contacts within the
Army, Navy, Air Force, and Department of Energy were also
queried for information about ground-water contamination
sites on federal facilities. Additional sites were
identified through professional contacts with environmental
consultants, and reports in the open technical literature.
In Chapter 2, statistical descriptions of the results of
this data collection effort are presented. The complete
collection of database reports resulting from the effort is
bound separately as Volume 3 of this report.
1.2.2 Collection of Detailed Site Data
The second data collection component involved the
acquisition of detailed information on selected sites where
ground-water extraction systems have been in operation long
enough to produce a record of system performance. In this
component of the study it was possible to draw at least
tentative conclusions about the effectiveness of the
extraction systems and the -site-specific factors affecting
them. On the basis of the information obtained a series of
case studies has been prepared. These case studies are
bound separately as Volume 2 of this report.
1-3
-------
The detailed data collected for each site included
background hydrogeologic information, the nature and history
of onsite activities, the distribution and characteristics
of ground-water contaminants, the design and operation of
the extraction system, and progress reports on the
effectiveness of the system. For those sites in the CERCLA
program, the information was usually presented in RI/FS
documents, remedial design reports, and quarterly, semi-
annual, or annual monitoring reports on system operations.
For RCRA sites, the information was often contained in
Part B permit applications, and annual post-closure
monitoring reports. For sites regulated by state agencies
the information was presented in a variety of formats and
was sometimes less than comprehensive.
One of the case studies included in Volume 2, for the Ville
Mercier site in Quebec, was taken directly from the
published proceedings of a technical conference. It is not
the product of this study but has been included because of
its applicability to the problems being investigated.
1.3 LIMITATIONS ON DATA COLLECTION
This study does not provide comprehensive coverage of all
contamination sites where ground-water extraction is planned
or in use. Because of time and resource limitations the
general data collection effort had to rely heavily on the
voluntary cooperation of individuals in state and federal
environmental agencies. Responses to requests for
assistance and data naturally varied from office to office
and not all geographical or administrative divisions are
covered equally. None-the-less, a representative cross-
section of ground-water extraction sites seems to have been
obtained.
Similarly, the sites selected for detailed case study do not
include all of the operating extraction systems. Most of
the sites identified in the general data collection effort
were still in the pre-implementation phase, but of those
systems that have begun operation, many have been in
operation for such a short time that no substantial
performance record has yet been generated. There are also
some sites where ground-water extraction has been going on
for a relatively long time, but the data collected is not
sufficiently detailed for use in a case study.
WDCR437/024.50
1-4
-------
Chapter 2
OVERVIEW OF RESULTS
2.1 GENERAL SITE DATA COLLECTION
The general site data collection effort obtained information
on 112 sites where ground water extraction has been chosen
as the remediation technique. This information was entered
into a database using dBase III Plus software. The variety
of sources from which the information was obtained caused
some inconsistency. Data were entered as objectively as
possible but some interpretation was necessary, for
instance, when a narrative description was provided that did
not fit neatly into the pre-selected database categories.
Additional information on any of the sites may be available
from the contact persons listed on the database reports in
Volume 3.
A concise summary of the data collected for the 112 sites
identified is presented in Table 2.1. The subsections that
follow will discuss and present statistics on the data
categories represented by the columns of the table.
2.1.1 Site Locations
Sites from all 10 EPA Regions are represented in the general
data. In addition to sites located within the United States,
data from one site in Canada and two sites in Puerto Rico
were entered into the database. The distribution of the
number of sites from each region is:
EPA Number
Region of Sites Percent
165
II 13 12
III 8 7
IV 44 39
V 10 9
VI 10 9
VII 1 1
VIII 1 1
IX 6 5
X 12 11
Other 1 1
112
2-1
-------
Table 2.1
SUMMARY OF GENERAL SITE DATA
SITE NAME
25TH STREET WELLFIELD
AIR PRODUCTS & CHEMICALS
AIRCO
AMERICAN CYANAMID
AMERICAN CYANAMID
AMERICAN WOOD TREATING
AMPHENOL
ANDERSON CORPORATION
ANNISTON ARMY DEPOT
AREA D - MCCLELLAN AFB
BALDWIN POLE & PILING
BAYOU BONFOUCA
BERKS SAND PIT
BF GOODRICH
BFl/CECOS
BLACK AND DECKER
BLOSENSKI LANDFILL
BOEING OF PORTLAND
BTL SPECIALTY RESINS CORP.
BURLINGTON NORTHERN RR
CAPE FEAR WOOD PRESERVING
CAVENHAM CORPORATION
CAVENHAM CORPORATION
CHEMTRONICS SITE
CIBA-GEIGY
COLEMAN-EVANS WOOD PRESERV.
CONOCO RERNING
COOPER BIOMEDICAL, INC.
DES MOINES TCE
DISTLER BRICKYARD
DISTLER FARM
DRAKE CHEMICAL
DUPONT-AXIS
ELECTRONIC CONTROLS, CORP.
ELECTRONIC INDUSTRIES
EMERSON ELECTRIC
ENSCO
EVANITE BATTERY SEPARATOR
STATE
FL
NJ
WV
WV
FL
MS
NY
MN
AL
CA
AL
LA
PA
KY
LA
NY
PA
OR
NY
MN
NC
MS
AL
NC
AL
FL
OK
NJ
IA
KY
KY
PA
AL
MA
MN
FL
AR
OR
EPA
REGION
IV
II
III
III
IV
IV
II
V
IV
IX
IV
VI
III
IV
VI
II
III
X
II
V
IV
IV
IV
IV
IV
IV
VI
II
VII
IV
IV
III
IV
1
V
IV
VI
X
Administrative
Program
T3
c
1
w
X
X
X
X
X
X
X
X
X
X
X
X
5
u
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
n
State Le
X
X
X
X
X
X
X
X
X
•
c
c.
c
X
X
«
5
X
X
Remedial
Objective
|
i
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
^
o>
1 Containi
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
c
cu
E
I
f
5:
X
X
X
sr
0>
>
§
r
L
2
X
X
X
X
X
X
X
r
c
3
"n
§
a
Exlractio
Type
CO
I
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
CO
c
o
X.
oi
5
X
X
5
o
X
X
System
Enhancement
O)
c
Q-
Q_
•§
ja
Q.
X
.2
I
1
X
X
X
Site Type
"O j£
§ f
-§ 3
D 55
o
%
CO
> £
N E
X
x
X
X
X
c
re
0.
1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
i
ii
X
1
0
X
X
X
X
Sontaminan
Type
= •
r
O
x
x
x
x
x
x
x
x
x
x
(
<7J
x
x
x
x
x
X
x
x
X
x
x
x
X
x
x
X
1
x
x
X
x
x
x
X
x
x
X
x
x
x
x
x
x
x
x
x
x
X
x
x
(
(
x
X
i!
a
5
x
x
x
x
x
x
x
c
X
x
X
x
o
D
1
5
2
1_
X
x
x
x
x
X
-------
Table 2.1
SUMMARY OF GENERAL SITE DATA
(CONTD)
SITE NAME
FAIRCHILD SEMICONDUCTOR
FERNWOOD WOOD TREATING CO
FMC CORPORATION
FMC CORPORATION
FORD MOTOR CO
FRONTIER HARD CHROME
FRUEHAUF CORPORATION
GEIGER/C&M OIL
GENERAL ELECTRIC/EAST ST.
GENERAL ELECTRIC/PINNELAS
GENERAL MILLS
GENRAD CORPORATION
GH.SON ROADS
GREAT LAKES CHEMICAL CORP.
GROVELAND WELLS
HARRIS CORPORATION
HELEVA LANDRLL
HOECHST/CELANESE NPL SITE
IBM - DAYTON
IBM • SAN JOSE
INDIAN BEND WASH
INTERNATIONAL PAPER CO.
KURT MANFACTURING
M-AREA SAVANNAH RIVER PLT
MEPCO/ELECTRA
MILLCREEK
MOBAY CORPORATION
MOBAY CORPORATION
NATIONAL STARCH & CHEM.
NICHOLS ENGINEERING
NORTHERN TELECOM
NORTHSIDE LANDFILL
OLIN CHEMICAL
OLIN CORP
PALMETTO WOOD
PENSACOLA NAVAL AIR STA.
PERDIDO
PERMAPOST PRODUCTS CO.
STATE
CA
MS
NY
MN
NJ
WA
AL
SC
MA
FL
MN
MA
MA
AR
MA
FL
PA
NC
NJ
CA
AZ
MS
MN
SC
NJ
PA
TX
WV
NC
NJ
FL
WA
KY
AL
SC
FL
AL
OR
EPA
REGION
IX
IV
N
V
n
X
IV
IV
i
IV
V
1
1
VI
1
IV
111
IV
n
IX
IX
IV
V
IV
n
in
VI
in
IV
n
IV
X
IV
IV
IV
IV
IV
X
Administrative
Program
j
X
X
X
X
X
X
X
X
X
X
X
X
X
X
£
X
X
X
X
X
X
X
X
X
X
X
X
State Lead 1
X
X
X
X
X
X
X
X
X
X
X
X
X
DOD / DOE* 1
*
*
X
1
6
X
Remedial
Objective
Restoration 1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Containment 1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
NAPL Recovery ]
X
X
X
3
o
1
X
:xtracNon
IT/P*
i
I
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1
s
L
i
i
2
1
X
X
X
X
System
Enhancements
Pulsed Pumping
X
X
X
Reinjectkx)
X
X
Site Type
[1
i!
5sj
j
ll
x
X
X
X
t
i
L
I
2
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
I
!t
< 0
i
=
X
X
X
X
/ont&min&nt
Type
5
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
i
X
X
X
X
X
X
X
X
X
X
i
I
5>
f
X
1 Low sorption
X
X
X
x
NAPLS
1 Floaters
X
X
X
Sinkers
X
X
X
X
X
X
X
X
X
AquHer Materials
O
X
X
X
X
X
X
X
X
X
X
§
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
\
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1
§
X
X
X
X
X
X
X
X
X
X
X
i
)
>
i
X
X
X
X
X
X
Iineroeooea
Sediments
X
X
X
i
j
X
X
X
1
J
X
X
X
X
i!
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
Table 2.1
SUMMARY OF GENERAL SITE DATA
(CONT'D)
SITE NAME
PHOENIX-GOODYEAR AIRPORT
PONDERS CORNER
REICHOLD CHEMICAL
REICHOLD PENSACOLA
ROCKY MOUNTAIN ARSENAL
ROLLINS-BATON ROUGE
SAPP BATTERY
SCRDI/DIXIANA NPL SITE
SEYMOUR RECYCLING
SIDNEY MINE
SITE A
SODYECO SITE
SPARTAN TECHNOLOGY
STAUFFER - ICI
TEXAS EASTMAN
TEXAS ECOLOGISTS
TIME OIL SITE
TOMAH PRODUCTS
TOWER CHEMICAL
TUSCON INT'L AIRPORT AREA
UN-NAMED
UNION CARBIDE, PONCE
UNITED CHROME PRODUCTS
UPJOHN
UTAH POWER & LIGHT CO.
VERONA WELL FIELD
VILLE MERCIER, QUEBEC
W.R. GRACE
WAMCHEM SITE
WESTERN PROCESSING
WHITT ACKER SITE
WILSON CORNERS (KSC)
WOOD TREATING, INC.
WURTSMITH AFB
WYCKOFF EAGLE HARBOR
ZELLWOOD
STATE
AZ
WA
WA
FL
CO
LA
FL
SC
IN
FL
FL
NC
NM
AL
TX
TX
WA
NJ
FL
AZ
FL
PR
OR
PR
ID
Ml
CN
MA
SC
WA
MN
FL
MS
Ml
WA
FL
EPA
REGION
IX
X
X
IV
VIII
VI
IV
IV
V
IV
IV
IV
VI
IV
VI
VI
X
II
IV
IX
IV
II
X
II
X
V
1
IV
X
V
IV
IV
V
X
IV
Administrative
Program
iji
!§«8i
W (T W Q O
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Remedial
Objective
| Restoration |
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
| Containment |
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
i W.M. ireaiment
X
X
X
£
X
X
X
X
X
| Leachate Coll.
X
X
Extraction
Type
01
5
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1 Well Points
X
X
X
1 Trenches
X
X
X
System
Enhancements
1 Pulsed Pumping
X
X
X
X
X
X
Reinjection
X
X
X
Site Type
1 Underground
1 Stroage Tank
X
D
I
sj E
IB
X
X
X
X
X
X
C
ra
Q.
1
»
1
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
1 Ammunition
1 Dump
-------
The large number of sites in Region IV should not be
interpreted to indicate that Region IV has more ground water
contamination problems than other regions. Instead, the
high percentage of Region IV sites shows that the
individuals familiar with sites in this region were more
responsive to the data gathering effort.
2.1.2 Administrative Programs
Five administrative programs are listed by name in Table
2.1, together with an addition column for other programs.
Seventy-eight percent of the 112 sites are under the Super-
fund or RCRA programs. The number of the sites in each
program can be broken down as:
EPA Number
Region of Sites Percent
Superfund 45 40
RCRA 42 38
State lead 24 21
DOD 5 4
DOE 2 2
Other 6 5
Some of the sites are associated with more than one
administrative program. The combination of Superfund and
RCRA occurred six times and the combination of RCRA and
State Lead occurred six times. In some cases, sites may
have been transferred from one program to another. In
others, a site may have multiple components that are
administered by different programs.
2.1.3 Remediation Obj ectives
The objectives of the ground-water extraction were
categorized as: (1) aquifer restoration, (2) plume
containment, (3) leachate collection, (4) well-head
treatment, and (5) nonaqueous liquid recovery. The number
of sites with each objective can be summarized as follows:
Number Percent of
Objective of Sites 112 Sites
Aquifer restoration 90 80
Plume containment 65 58
Well head treatment 8 7
Nonaqueous liquid recovery 16 14
Leachate collection 3 3
2-2
-------
The most common objectives were aquifer restoration and
plume containment. Ninety-seven percent of the sites claim
either one or the other to be their objective. Forty-one
percent (46 sites) have both options as their objective. It
is difficult to judge from the general data whether or not
aquifer restoration is feasible at each of the sites where
it is listed as a goal.
2.1.4 Type of Extraction System
Three extraction system options identified from the general
data were drilled wells, well points, and trenches. Ninety-
two percent of the sites use drilled wells in their
extraction systems. The number of sites in each category
is:
Number Percent of
Extraction System of Sites 112 Sites
Drilled wells 103 92
Well points 5 4
Trenches 9 8
Unknown 3 3
At one of the sites (Sidney Mines) all three extraction
types are being employed.
2.1.5 Extraction System Enhancements
Pulsed pumping is employed as a system enhancement at ten
sites and reinjection is used at eight sites. Other
remedial technologies such as site capping, excavation,
vacuum extraction, and bioreclamation, which are sometimes
used in conjunction with ground-water extraction systems,
were not included in the general data collection effort.
2.1.6 Type of Site
The types of sites encountered in the general data
collection effort are categorized as: Underground Storage
Tank (UST), Hazardous Waste Dump, Industrial Plant,
2-3
-------
Ammunition Dump, and Other. The distribution of the sites
in the database for these categories is:
Number Percent of
Site Type of Sites 112 Sites
UST 1 1
Hazardous waste dump 15 13
Ammunition dump 1 1
Industrial plant 81 72
Other 15 13
2.1.7 Type of Contaminant
Four categories of contaminants were identified in the data
collection effort. They were high sorption, low sorption,
organics, and metals. In most cases, the individuals
providing the data did not indicate whether the contaminants
had high or low sorption potential. The number of sites in
each category is as follows:
Number Percent of
Contaminant Type of Sites 112 Sites
High sorption 13 12
Low sorption 26 23
Both high and low 8 7
Organics 106 95
Metals 32 29
Both organics and metals 28 25
2.1.8 Presence of Nonaqueous Liquids
Nonaqueous phase liquids (NAPLs) are present in 46 of the
sites in the general survey. Of these 46, 37 sites have
NAPLs that are more dense than water and 15 sites have NAPLs
that are less dense than water. Six sites have both types
of NAPLs.
2.1.9 Geologic Materials
Geologic materials identified in the general data collection
program include: clay, silt, sand, gravel, fractured rock,
interbedded sediments, and limestone. Most respondents
circled multiple materials since most sites are underlain by
2-4
-------
several types of geologic materials. The number of sites
having materials in each category is:
Number Percent of
Geologic Material of Sites 112 Sites
Clay 30 27
Silt 46 41
Sand 88 79
Gravel 30 27
Fractured Rock 15 13
Interbedded Sediments 16 14
Limestone 6 5
Other 17 15
2.1.10 Implementation Status
Forty-seven of the sites (42 percent) have pumping in
progress. These sites are indicated in the last column of
Table 2.1. All other sites are in a pre-implementation
phase.
2.2 DETAILED DATA COLLECTION AND CASE STUDIES
Volume 2 of this report consists of case studies for
19 hazardous waste sites at which ground-water extraction
systems have been in operation long enough to produce a
record of performance. Of these case studies, 18 were
developed as part of this project and one, the Ville Mercier
study, was taken directly from the technical literature.
The case studies describe the site history, geology,
hydrogeology, and waste characteristics to give context to
the problem at each site. The remediation system and its
performance history is then described, followed by a summary
of the case study. Most of the statements concerning site
conditions and extraction system performance are derived
from information provided by the parties responsible for the
sites. However, some conclusions have also been drawn as
part of this investigation by the individuals who prepared
the case studies. These conclusions are confined to the
summary or conclusions sections of the case studies.
Table 2.2 is a summary of the site characteristics at the
19 case study sites. The geographic distribution of case
study sites is shown in Figure 2.1. The longest operating
record (15 years) is available for the Olin Corporation
site, where ground-water extraction was started in 1974. The
2-5
-------
Table 2.2
SIMUKY OF CASE STUDY SITE CHARACTERISTICS
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Site Name
Amphenol Corporation
Black & Decker
Dee Koines TCE
Dupont Mobile Plant
Emerson Electric Company
Fairchild Semiconductor
General Mills, Inc.
GenRad Corporation
Harris Corporation
IBM Dayton
IBM San Jose
Nichols Engineering
Olln Corporation
Ponders Corner
Savannah River Plant
Site A
Utah Power & Light
Date of Initial
Extraction
January 1987s
May 1988a
December 1987a
December 1985a
December 1984b
1982a
Late 1985a
Late 1987a
April 1984a
March 1978a
May 1982a
January 1 988a
1974"
September 1984a
September 1985a
August 1988a
October 1985a
Remedial
Objective
Restoration
Restoration*
Restoration'
Containment
Restoration*
Containment
Restoration^
Restoration'
Well-heated
treatment &
Restoration
Was restoration,
now containment
Restoration
Restoration*
Containment
Well-head
Treatment
Mass reduction
Restoration1
Containment
Chemicals
Present
Organics
Organics
Organics
Organics
Organics
Organics
Organics
Low Sorption
Organics
Organics
Organics
Organics
Organics
Organics
Low Sorption
Organics
Low Sorption
Organics
Organics
Organics
NAPLs
Present
No
No
No
No
No
Maybe
Maybe
No
No
Yes
Yes
Maybe
No
No
No
No
Yes
Geologic Environment
Unconsolldated glacio-
fluvlal sediments
Glacial till & fractured
sandstone
Unconsolldated glaclo-
fluvlal sediments
Alluvial sand 4 clay
Sand
Alluvial sand & gravel
with silt & clay layers
Peat, glacial deposits.
& fractured rock
Glacial sand, gravel
Sand & shell with a
clay layer
Sand with clay layers
Alluvial sand & gravel
with silt & clay layers
Weathered & fractaured
shale
Unconsolidated glaclo-
fluvial sediments
Unconsolldated glacio-
fluvial sediments
Coastal plain sand,
silt i clay layers
Limestone & sand
Alluvium & fractured
basalt
Innovative
Technologies
Fracture
Enhancement
Slurry wall
Intermittent
pumping
Well points
Well points
Re Inject Ion
Vapor extraction
Intermittent
pumping
Admlnlst rat ive
Program
RCRA
RCRA
Superfund
RCRA
State Lead
State Lead
State Lead
RCRA
Superfund &
State Lead
State Lead
State Lead
State Lead
State Lead
Superfund
DOE
Superfund
RCRA
-------
Table 2.2 (Continued)
SIMUKY OF CASE STUDY SITE CHARACTERISTICS
Site
No. Sit- Name
18 Verona Well Field
19 Vlllle Herder
Date of Initial
Extraction
May 1984a
1983s
Remedial
Objective
Restoration1 &
Containment
Containment
Chemicals
Present
Organics
High and Low
Sorptlon
Organics
NAPLs
Present
Yes
Yes
Geologic Environment
Glacial sand, gravel,
& clay
Unconsolldated glacial
sediments & fractured
rock
Innovative
Technologies
Vapor extraction
Administrative
Program
Superf und
Province of
Quebec
Notes;
a Extraction still In progress.
k Remediation completed and extraction system shut down in July 1987.
1 Restoration to concentration goals equal to or less than health-based standards--MCLs or 10~6 excess cancer risk concentrations.
^ Restoration to site-specific goals not directly related to health-based standards.
WDCR437/040.50
-------
W DC 61621.AO 02
Engineering
IBM Dayton
\ r
Falrchlld Semiconductor
Savannah River Plant
Emerson Electric
Harris Corp.
Site A
Figure 2.1
GEOGRAPHIC DISTRIBUTION OF CASE STUDY SITES
-------
site with the shortest record of operation is Site A, where
extraction began at the end of August 1988. Table 2.2 lists
the approximate starting dates of the extraction systems
described in the case studies. The Emerson Electric site is
the only one of the 19 at which the remediation has been
completed and extraction has been terminated.
Table 2.2 also shows the administrative program under which
the ground-water extraction systems are operated. There are
5 Superfund sites, 5 RCRA sites, 8 state lead sites, 1 DOE
site, and 1 site administered by the province of Quebec,
Canada. The Harris Corporation site includes three
administratively separate parts. All three are administered
by the Florida Department of Environmental Regulation, but
one is also a Superfund site. Consequently, this site is
counted as both Superfund and state lead.
2.2.1 Remediation Obj ectives
In this study it is important to recognize the differences
in remedial objectives for the extraction systems described
in the case studies. The remedial goals generally have an
important influence on the design and operation of the
systems. They must certainly be taken into account when
evaluating system effectiveness.
The primary objectives of the ground-water extraction
systems at the case study sites are listed in Table 2.2. At
12 of the 19 sites, the current objective includes aquifer
restoration. This generally means that the contaminated
aquifer is to be remediated to a specified concentration
goal. The cleanup standard used is not always a health- or
environment-based criterion as defined by current regula-
tions. In some cases, site-specific cleanup goals were
established. In others, state requirements that are more
stringent than the federal standards have been adopted.
At the Savannah River Plant site, the remediation goal is to
remove 99 percent of the estimated mass of subsurface
contaminants. This is not a health-based criterion, but it
is a quantitative goal for removal of contaminants.
At the IBM site in Dayton, New Jersey, the initial remedial
objective was aquifer restoration. However, because it is
thought that nonaqueous phase liquids (NAPLs) are present in
the aquifer, that goal has been abandoned in favor of plume
containment. The design and operating procedures of the
extraction system were then modified to increase its
containment efficiency. For this reason, both restoration
and containment are listed as objectives for this site.
2-6
-------
Both restoration and containment are also listed as the
objectives for the Verona Well Field site. This was done
because the site includes two extraction systems. The
objective of one is plume containment, and the other is
intended to restore aquifer quality.
Similarly, the Harris Corporation site has multiple
extraction systems with different objectives. One is a
well-head treatment system, which permits continued
operation of water-supply wells in a contaminated aquifer.
The other is an aquifer restoration system located near the
source of the contamination.
Five of the sites listed in Table 2.2 have been designated
as containment systems. This means that aquifer restoration
is not expected within the foreseeable future at these
sites. This is a somewhat restricted definition of the
containment objective, because plume containment is also an
objective at most aquifer remediation sites. However, when
both restoration and containment are stated objectives of
the remediation, aquifer restoration is considered to be the
primary objective in this study.
Two of the sites studied—Harris Corporation and Ponders
Corner—have well-head treatment systems. These are sites
where the primary objective of ground-water extraction is to
provide continued water supply from a contaminated aquifer.
In both of these cases, the source of the contamination is
known, but this is not true at all well-head treatment
sites. At the Harris Corp. site the contaminant source is
being remediated by a separate ground-water extraction
system in addition to the well-head treatment system. At
Ponders Corner it is expected that the operation of the
production wells with well-head treatment will continue
indefinitely. This should eventually result in aquifer
restoration. A minimum restoration period of 10 years has
been estimated, but no maximum has been projected.
The Verona Well Field is not listed as a well-head treatment
site even though it is a producing well field, because the
contaminated ground water produced from the blocking wells
is not used for water supply. The wells that intercept the
contaminant plume are used as blocking wells, which makes
this a containment system rather than well-head treatment.
2.2.2 Contaminant Characteristics
Most of the case study sites involved contamination with
volatile organic compounds. The most frequently encountered
contaminants were chlorinated ethanes and ethenes. The
2-7
-------
Dupont Mobile site involved pesticides in addition to
chlorinated organics. However, because of their low
mobility the pesticides did not figure prominently in the
evaluation of extraction system performance. At the IBM San
Jose site, Freon 113 was an important contaminant in
addition to chlorinated ethenes and ethanes. At the Utah
Power and Light site, the only contaminant of concern was
creosote, a mixture consisting primarily of phenolics and
polycyclic aromatic hydrocarbons. Contamination at the
Ville Mercier site included a wide variety of organic
compounds and some metals.
The case studies were not intentionally limited to organic
contaminants. Several sites were identified where ground-
water extraction was being used to remediate aquifers
contaminated with metals. However, case studies could not
be done for them because of insufficient performance data.
The ground-water remediation programs at several of the
sites studied were complicated by the presence of
contaminants in nonaqueous phase, as indicated in Table 2.2.
At five of the sites, it has been concluded during site
investigations that NAPLs are present. This is indicated by
a "yes" in the NAPL column of Table 2.2. At three other
sites, the presence of NAPLs is suspected by the authors of
this report, either because of the high concentrations of
dissolved contaminants reported or because of the nature of
the contamination source. These sites are designated by a
"maybe" in Table 2.2. In many cases no special effort has
been made in the site investigations to determine whether
NAPLs are present or not.
2.2.3 Geologic Environments
Five of the case studies deal with contamination of
fractured rock aquifers. These sites are: Nichols
Engineering, Utah Power & Light, General Mills, Black &
Decker, and Ville Mercier. Fractured rock aquifers are of
special interest because ground-water flow and contaminant
transport in them is potentially more difficult to analyze
and control. The sites with contamination primarily in
unconsolidated sediments include a variety of geologic
types, as described in Table 2.2.
At five of the sites, ground water is extracted from more
than one aquifer. This generally requires special
coordination of the pumping rates in the different groups of
wells to properly manage the vertical potentiometric
2-8
-------
gradients. These multi-aquifer sites are: Fairchild
Semiconductor, IBM Dayton, IBM San Jose, General Mills, and
Utah Power & Light.
2.2.4 Use of Innovative Technologies
Few of the sites covered by case studies employed any
innovative technologies to enhance the performance of the
ground-water extraction systems. All of the sites used
drilled wells, but at two sites, well points were also used.
Well points are wells that are driven or jetted directly
into the ground rather than installed in a drilled hole.
Well points are often shallow and are often pumped by
surface vacuum pumps.
At the Harris Corporation site, a line of well points was
installed to supplement the ground-water extraction from
drilled wells. However, the well points developed
operational problems and were soon replaced with two
conventional drilled wells. At the IBM Dayton site, two
lines of well points were installed to supplement the
conventional extraction well system. One line of well
points was used for ground-water withdrawal, and the other
was used for reinjection.
The IBM Dayton site was the only case study where reinjec-
tion of treated ground water was employed. In this case,
the purpose of reinjection was to influence ground-water
flow patterns. The injection wells were operated for only a
few months before they had to be abandoned due to clogging
of the formation. At the Utah Power & Light site, reinjec-
tion was considered as an initial design option, but was
rejected because it would interfere with the control of
vertical gradients between aquifers.
At the Fairchild Semiconductor site the onsite extraction
system was supplemented by an encircling slurry wall to
isolate the most heavily contaminated part of the surficial
aquifer. The slurry wall has helped the extraction system
establish the desired hydraulic gradients between aquifers.
Vapor extraction has been used to accelerate cleanup of
contamination above the water table at the Ponders Corner
and Verona Well Field sites.
At the Black & Decker site, the density of bedrock
fracturing was enhanced by the use of explosives to increase
the effectiveness of the extraction wells in capturing
contaminants.
2-9
-------
Intermittent pumping of the extraction wells has been
practiced at the Genrad and Utah Power & Light sites.
However, this was not intentionally done to increase
extraction efficiency in either case. At Genrad, the
extraction has been intermittent to avoid freezing of the
above-ground piping during the winter months. At Utah Power
& Light, pumping of individual wells has been intermittent
because of the limited productivity of the aquifers and
frequent shutdowns of the treatment plant. At this site,
the relatively high ground-water velocities associated with
restarting individual wells may have had a beneficial effect
on the recovery of nonaqueous phase liquids, but this has
not been clearly demonstrated.
2.2.5 Extraction System Design Information
Table 2.3 presents a summary of information concerning the
design of the extraction systems at the 19 case study sites.
The first several columns in the table indicate the kinds of
site data and the analytical methods that were used in
designing the extraction systems. The remaining columns
indicate the extraction capacity of the systems, whether the
system configuration has been adjusted in response to
operating experience, and (where applicable) the expected
time required to complete the remediation. This information
was taken from the design and operating reports collected
for the case study sites during the detailed data collection
phase of the study. For some of the sites, the available
reports may not have provided complete descriptions of the
site data collected or the design methods used.
At all but two of the sites, aquifer testing was done to
help determine the hydrogeologic characteristics of the
aquifer to be remediated. Aquifer testing, as referred to
here, means that a well on the site was pumped at a
controlled rate for a certain period of the time, and the
resulting water level drawdown was measured in separate
observation wells.
Numerical modeling was used to analyze ground-water flow at
most of the case study sites. The flow models were
generally used to determine capture zones of the planned
extraction wells. At six of the sites, the flow patterns
predicted by the models were used to estimate the travel
times required for contaminants at the edge of the
contaminant plume to reach the extraction wells. However,
numerical contaminant transport modeling was not used at any
of the sites.
2-10
-------
Table 2.3
SUMMARY OF DESIGN INFORMATION FOR CASE STUDY SITES
Site
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Notes
Aquifer Flow Travel Time
Site Name
Amphenol Corporation
Black & Decker, Inc.
Des Moines TCE
Du Pont Mobile Plant
Emerson Electric Co.
Fairchild Semiconductor
General Mills, Inc.
GenRad Corporation
Harris Corporation
IBM Dayton
IBM San Jose
Nichols Engineering
Olin Corporation
Ponders Corner
Savannah River Plant
Site A
Utah Power & Light
Verona Well Field
Ville Mercierc
Tests
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
--
Model
Yes
No
Yes
No
Yes
No
Yes
Yes
Yes
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
--
Analysis
Yes
No
Yes
No
Yes
No
No
No
No
Yes
No
No
No
No
Yes
Yes
No
No
--
Multilevel Soil
Sampling
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Sampling
Yes
No
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
--
Sorption
Considered
Yes
No
Yes
Yes
Yes
No
No
Yes
No
Yes
Yes
Yes
No
Yes
Yes
No
Yes
Yes
Yes
'Projection of cleanup period is not applicable to containment systems.
"Not explicitly projected; figure based on interpretation of design information.
Increase in extraction rates is planned.
°Little design information is given in the case study for this site.
Containment or well-head treatment portion.
Restoration portion.
Pilot
Test ini
No
No
No
No
No
No
No
Yes
No
Yes
No
Yes
No
No
No
No
Yes
No
System Max. Pumping Projected
Modifications Rate (gpm) Cleanup Period
No
No
No
Yes
No
Yes
No
No
Yes
Yes
Yes
Yes
No
No
No"
Yes
Yes
Yes
No
200
15-20
1300
150-180
30
9,200
370
30
l.OOO1
3002
1,000
6,000
65
2,600-3,600
1,200
440
50
200
2.0001
4002
750
5 to 10 years
Not Projected
Not Projected
N/A*
9 Months
N/A*
Not Projected
> 5 Years
Not Projected
6-11 Years
10 Years*
2.25 Years
N/A*
> 10 Years
30 Years
25-60 Days
N/A*
Not Projected
N/A*
WDCM09/014.50
-------
At all but two of the sites, the subsurface exploration
program included multi-level ground-water sampling to
determine the vertical distribution of contaminants. This
was done at all of the sites involving more than one aquifer
and also at many of the single-aquifer sites.
Subsurface soil samples were taken and analyzed at all but
one of the case study sites. In most cases, soil sampling
was conducted for the purpose of identifying the
contaminants present and characterizing the contaminant
source. Only rarely were soils sampled over a wider area to
study the contaminant sorption characteristics of the
aquifer or to search for contaminants in the nonaqueous
phase. However, sorption was considered, in one way or
another, at all but six of the sites.
Pilot testing, as referred to in Table 2.3, means that a
small-scale extraction system was initially installed to
provide data for the design of the final system. This was
done at four of the case study sites. System modification
means that the configuration of the full-scale system was
adjusted on the basis of information gathered during
performance monitoring. This happened at about half of the
sites. In some cases, wells were added or pumping rates
were increased because monitoring showed that the initial
system was inadequate. In other cases, extraction at some
wells was terminated as a result of the progress of the
remediation.
The last column of Table 2.3 lists the projected cleanup
period as estimated by the designers of the extraction
system. At sites where the remedial objective is not
aquifer restoration, such a projection is not applicable,
and the extraction system is expected to continue operating
indefinitely. Even for several of the aquifer restoration
sites, no projection of the cleanup time has been made.
2.3 GENERAL CONCLUSIONS
The following general conclusions can be drawn from the case
study data collected in this project:
1. The ground-water extraction systems were generally
effective in maintaining hydraulic containment of
contaminant plumes, thus preventing further migration
of contaminants. Even so, the design of successful
containment systems requires careful study of site
hydrogeology. Hydraulic containment of the contaminant
plume was a goal, either primary or secondary, at all
2-11
-------
of the 19 case study sites. In most cases, successful
containment has been demonstrated.
2. Significant removal of contaminant mass from the
subsurface can often be achieved by the operation of
ground-water extraction systems. Where site conditions
are favorable, and the extraction system is properly
designed and operated, it may be possible to remediate
the aquifer to health-based levels. However, the time
required for complete remediation is usually longer
than initially estimated.
3. The contaminant concentrations observed at sites where
aquifer restoration systems are in operation typically
show a rapid initial decrease and then level off or
decrease at a greatly reduced rate. This effect is to
be expected because the mass reductions are greatest
when contaminants are being removed at high concentra-
tion. This effect may also be caused in part by
dilution due to the unintentional withdrawal of ground
water from regions outside the contaminant plume. Care
must be taken to design the aquifer restoration system
in a way that maximizes extraction of the most highly
contaminated ground water.
4. The success of ground-water extraction systems is
highly dependent on site hydrogeology and contaminant
characteristics. Site conditions must be taken into
account in the selection of realistic system objec-
tives. Even when appropriate objectives have been
chosen, the extraction system may be inefficient in
achieving them if its design is not based on an
adequate site investigation. The data collection
practices at some of the sites reviewed did not provide
enough information to permit thorough characterization
of the subsurface and of contaminant interactions with
it.
5. Well-head treatment systems can provide plume contain-
ment while maintaining water supply from contaminated
aquifers. They are usually installed as a cost-
effective remedy for water supply needs. However, at
some of the sites studied they were integrated into the
overall aquifer remediation as barrier systems to
prevent uncontrolled plume migration.
6. It is important to consider the contaminated sub-
surface, both saturated and unsaturated zones, as an
integrated whole, and to design remediation programs to
address source areas and unsaturated zone contamination
2-12
-------
in addition to the ground-water contamination in the
saturated part of the aquifer. Enhanced methods such
as pulsed pumping and vapor extraction can potentially
improve the effectiveness of ground-water extraction
systems by addressing residual contaminant sources in
the vadose zone.
2.3.1 Aquifer Restoration
Of the 19 case studies presented in this investigation, only
one covers a site where a complete and final aquifer
restoration is claimed. This is the Emerson Electric
Company site in Altamonte Springs, Florida. The contaminant
concentrations in the recovered ground water were reduced
from more than 100 ppb for several compounds, and more than
1,000 ppb for one, to less than the detection limits in
approximately 2-1/2 years. This cleanup was initially
expected to take only 7 months. The progress of the cleanup
was measured by taking monthly samples of the ground water
from the combined pumping of the five recovery wells. The
termination criteria that had been established for the site
were met when two successive sampling events showed concen-
trations below the cleanup levels. The system was then
turned off. When two rounds of post-termination sampling
from the extraction wells showed concentrations still below
cleanup levels, the site was recommended for deletion from
the list of state action sites.
Several of the other case studies cover sites where the
extraction system appears to be progressing toward eventual
aquifer remediation. At Site A, the contaminant concentra-
tions have been reduced below the cleanup goals in most of
the monitoring wells. In two monitoring wells the concen-
trations were still slightly above the goal for benzene
after 199 days of extraction, but the trend in these wells
appears to be downward also. It was initially projected
that the cleanup would take only 60 days.
At the Amphenol site in Sidney, New York, only about two
years of performance record are available since the start of
extraction. However, the reduction in concentrations
observed in the monitoring wells appears to be consistent
with the initial prediction that aquifer restoration would
be completed in 5 to 10 years.
At the Fairchild Semiconductor site in San Jose, California,
aquifer restoration was not an expressed remedial objective,
but large offsite areas of three contaminated aquifers have
been cleaned up after seven years of extraction. Approxi-
mately, 90,000 pounds of chlorinated solvents have been
2-13
-------
removed during this time. There are still large areas of
offsite contamination, but these appear to be shrinking
steadily. Onsite contamination of the uppermost aquifer
includes contaminants in nonaqueous phase, which are being
contained by a combination of extraction wells and a slurry
wall. This onsite containment system will probably need to
be operated indefinitely because of the NAPLs.
For the other eight case study sites where aquifer restora-
tion is the objective, the progress that has been made
toward this goal varies. At Des Moines TCE, GenRad, and
Black & Decker, the data sets are either too short or too
sparse to permit a conclusion to be drawn. At Verona Well
Field, the existence of a floating layer of nonaqueous phase
contaminants makes it unlikely that complete restoration
will be achieved in the foreseeable future, even though
significant quantities of contaminants are being removed by
both the ground water and vapor extraction systems. At the
IBM site in San Jose, California, extraction has been
continuing for more than 6 years. Large areas of the
contaminated aquifers offsite have been cleaned up to MCLs
or 10"6 excess cancer risk concentrations, but reduction in
contaminant concentrations to the more stringent site goals
set by the state of California has generally been slow.
Onsite contamination has been persistent but some progress
is evident. At Nichols Engineering, the available perform-
ance record is short, but so far it does not indicate
convincing progress. At the General Mills site in
Minneapolis, Minnesota, the extraction systems have achieved
some initial concentration reductions. However, the
apparent existence of a residual contamination source and
the possible presence of NAPLs in the aquifers make it
doubtful that restoration will be completed soon.
At the IBM site in Dayton, New Jersey, the goal of the
aquifer remediation was thought to have been attained in
1984, after 6 years of groundwater extraction. This conclu-
sion was based partly on the reduction of concentrations to
low stable levels in most of the offsite wells, and partly
on the criteria that had been established for system shut-
down. These criteria included the option to terminate the
remediation if performance records showed that concentra-
tions were no longer declining. Continued monitoring after
extraction had been discontinued showed steady increases in
the contaminant concentrations in monitoring wells as the
original contaminant plume was re-established. Further
study indicated that the re-emergence of the plume may have
been caused by the presence of contaminants in nonaqueous
phase. Consequently, the objective of the remediation was
changed from aquifer restoration to plume containment.
2-14
-------
At the Savannah River Plant, the extraction system may not
have enough capacity to effectively deal with the magnitude
of the contamination problem. Parts of the plume are not
presently within the capture zone of the extraction wells.
However, an increase in extraction system capacity is
planned. Considerable quantities of contaminant have been
recovered, but the reduction in contaminant mass in the
aquifer appears to be much less than the mass of
contaminants recovered. This may be due to recontamination
of the aquifer by residual sources including contaminants
sorbed to the aquifer materials.
It should be noted that the two case study sites that appear
to be in the most advanced stages of aquifer restoration,
Emerson Electric and Site A, are not those with the most
comprehensive site investigations. The aquifer
contamination at both of these sites was thought to be
relatively small in scale, although the extent of the
contaminant plume has not been fully delineated in either
case. This, together with the relative simplicity of the
hydrogeologic situation at both sites, has probably
contributed to their apparent responsiveness to remediation.
However, the data gathered during site characterization and
performance monitoring at these sites are too limited to
permit this to be stated with unqualified confidence.
2.3.2 Migration Control
At most of the case study sites where the objective is plume
containment, the extraction systems seem to have attained
this goal. At Fairchild Semiconductor, Verona Well Field,
Utah Power & Light, Olin Corporation, and Ville Mercier, the
measured potentiometric head distributions indicate that the
contaminant plumes are within the capture zones of the
wells.
At the Dupont Mobile site, control of contaminant migration
may not have been completely established. Potentiometric
head measurements at this site appear to show complete
hydraulic containment of the contaminant plume. However,
contaminant mass balance calculations indicate that only
about half of the contamination approaching the extraction
wells is being recovered. The discrepancy may be due to
inaccuracy in the estimation method used to estimate the
mass flux of the contaminant plume. However, it could also
arise from incomplete characterization of the ground-water
flow patterns around the extraction wells. The ground-water
flow has been analyzed two-dimensionally in the horizontal
plane, with water levels being measured by shallow partially
penetrating wells. The extraction wells are also partially
2-15
-------
penetrating, and their capture zone may not extend all the
way to the bottom of the contaminant plume. This
illustrates the need for adequate vertical hydrogeologic
site characterization in the design of ground-water
extraction systems.
At the Harris Corporation site in Palm Bay, Florida the on-
site extraction system has not recovered contamination that
migrated offsite before the remediation system was activated.
However, this contamination is captured by the well-head
treatment system on the adjoining property. The onsite
extraction system does appear to be preventing the contamin-
ant plume beneath the Harris facility from spreading.
Incomplete migration control has been observed at the Savannah
River Plant, and the extraction capacity of the ground-water
remediation system is being increased to correct the problem.
2.3.3 Weil-Head Treatment
In a well-head treatment system, contaminants present in the
water extracted by the production wells are removed before the
water is put to its intended use. Success in well-head treat-
ment is generally easier to achieve than in aquifer remedia-
tion systems because it is not necessary to address the
complexities of subsurface contamination. The types and
estimated concentrations of contaminants must be estimated in
order to design the treatment system, but once this is done,
the success of the system depends on the treatment effective-
ness rather than on the effectiveness of contaminant recovery.
The need to install additional wells is reduced and the moni-
toring requirements are generally less demanding with well-
head treatment systems, making them less expensive to operate.
The underlying contamination problem is generally not
addressed directly, but some improvements in ground-water
quality often occur.
Well-head treatment was used at two sites for which case
studies were done—Ponders Corner and Harris Corporation. At
Ponders Corner, the well-head treatment system was installed
as an interim remedial action intended primarily to restore
water supply. However, because the affected production wells
are near the center of contamination, the operation of this
system has significantly improved aquifer quality compared to
initial levels. Well-head treatment can often be an important
interim remedial measure that allows the beneficial use of
ground water to continue during the formulation of a final
response. In cases where aquifer restoration is not practic-
able, well-head treatment can still be a realistic alternative
that allows the beneficial use of ground water to continue.
2-16
-------
At the Harris Corporation site, the water purveyor affected
by contamination from the Harris Corporation facilities has
installed a well-head treatment system to maintain its water
supply. This treatment system is financed by Harris Corpor-
ation and forms part of a complex containment and aquifer
restoration effort in operation at the Harris facilities.
This is a case in which well-head treatment is part of a
phased approach that permits beneficial use of the aquifer
in the short term while a long-term solution to its restor-
ation is being attempted.
WDCR161/004.50
2-17
-------
Chapter 3
FACTORS AFFECTING SYSTEM DESIGN AND PERFORMANCE
Ground-water extraction technology is based on two
fundamental assumptions. First, it is assumed that a well
system can produce ground-water flow patterns that will
permit the wells to withdraw all of the contaminated ground
water from the aquifer. Second, it is assumed that the
contaminants will come out of the aquifer with the water.
In an ideal hydrogeologic system with simple pre-existing
flow patterns, homogeneous aquifer properties, and with
mobile contaminants present only in aqueous solution,
ground-water extraction can work quite well. However, most
real-world sites are more complex than this, and most
departures from the ideal conditions described above tend to
reduce the potential effectiveness of ground-water
extraction. This chapter will discuss the effects of some
of these departures from ideal conditions on the design and
performance of extraction systems.
Again, it is necessary to distinguish the different
objectives of ground-water extraction systems. For
migration control systems it is only necessary to establish
a hydrodynamic regime that will ensure capture of the
contaminated ground water. The time required for
contaminant removal is not an issue. For aquifer
remediation systems, control must be established over the
movement of both the ground water and the contaminants.
Because migration control systems have fewer requirements,
they will be considered first.
3.1 FACTORS AFFECTING MIGRATION CONTROL
The design of successful migration control systems depends
on the ability to predict the ground-water flow patterns
that will be produced when the effects of the extraction
wells are added to the pre-existing flow system. Before
such predictions can be made, an understanding of the
hydrogeologic properties of the site and the spatial
distribution of the contamination must be acquired. Chapter
4 discusses the information requirements for system design.
3.1.1 Contaminant Distribution
The proper definition of the vertical and horizontal limits
of contamination is essential in designing almost all
remediation systems. The failure to define the limits of
3-1
-------
the plume can lead to incorrect recovery well placement and
pumping rates.
3.1.2 Effects of Pre-Existing Gradients
The potentiometric head distribution that exists in the
aquifer before the extraction well system is added is
referred to as the field of pre-existing gradients. This is
not necessarily the natural flow field, because the flow may
be affected by nearby ground-water withdrawals.
At sites where the pre-existing gradients are uniform and
unidirectional, the analysis of migration control wells is
relatively simple. A well-known technical article by
Javandel and Tsang (1986) provides a graphical method based
on potential flow theory for the design of single-well or
multiple-well plume containment systems in such cases. An
equivalent analysis can be performed by any trained hydro-
geologist simply by superimposing the drawdown predicted for
the extraction wells onto the uniform gradient field.
Strictly speaking, this method applies only to flow in
confined aquifers, but the error involved in applying it to
unconfined situations is generally small.
When the pre-existing flow field is nonuniform, the analysis
of the effects of extraction wells is likely to be more
complicated. If the nonuniformity of the flow is caused by
the presence of nearby production wells, as is often the
case near well fields, the superposition principle still
holds. The superposition procedure will then be somewhat
more cumbersome than in the case of a uniform regional
gradient, but it is still not particularly difficult.
However, the observation that the pre-existing flow field is
complex is often an indication of significant heterogeneity
in aquifer properties. This is considerably more difficult
to deal with, as discussed below.
Case study sites where the flow induced by nearby production
wells may have affected the extraction system include the
Dupont Mobile site and the Harris Corporation site.
In the above discussion of gradient fields, the problem has
been implicitly interpreted as being two-dimensional in the
horizontal plane. This is one of the common idealizations
used to simplify reality for ease of analysis. However, if
there are significant vertical gradients in the aquifer to
be remediated, a two-dimensional representation is not
adequate. Significant downward flow in an aquifer can
indicate that it is discharging to a lower hydrologic unit.
The extraction well system must then be capable of reversing
3-2
-------
the flow between aquifers in the contaminated area as well
as effecting plume capture with respect to horizontal flow.
Case study sites where flow between aquifers was of concern
include Savannah River Plant A/M-area, General Mills,
Fairchild Semiconductor, IBM San Jose, IBM Dayton, and Utah
Power & Light. Vertical flow may have been important at
Site A also, but it was not taken into account in the design
of the extraction well.
3.1.3 Effects of Aquifer Properties
3.1.3.1 Transmissivity
The aquifer*s response to ground-water withdrawals from the
extraction wells is determined primarily by its hydraulic
conductivity and saturated thickness. The product of these
two properties is the aquifer transmissivity.
In a low-transmissivity aquifer, the radius of influence of
an extraction well will be smaller than it would be in a
high-transmissivity aquifer. This is illustrated
schematically in Figure 3.1. Consequently, in an aquifer
with low transmissivity, more extraction wells must be used
and they must be more closely spaced if an effective
hydraulic barrier is to be formed. The spacing problem may
be further aggravated in unconfined aquifers, where the
aquifer's saturated thickness may limit the maximum drawdown
of individual wells. In aquifers with such restrictive
conditions, extraction trenches or lines of closely spaced
well points are sometimes used instead of drilled extraction
wells. Among the case study sites there were none that used
extraction trenches. However, as Table 2.4 shows, several
of the sites identified in the general data survey employ
trenches for ground-water extraction. Lines of well points
were used at IBM Dayton and at the Harris Corp. site, even
though the transmissivities of the aquifers there were not
particularly low.
3.1.3.2 Heterogeneity
A migration control system that would be effective in a
homogeneous aquifer may fail to completely capture the
contamination in an aquifer that has nonuniform hydro-
geologic properties. Heterogeneity can cause shifts in the
cones-of-depression around individual wells that may allow
contaminants to flow between or around them. Vertical
differences in hydraulic conductivity at an extraction well
can cause its radius of influence to vary with depth. This
could provide gaps for the escape of contaminants between
extraction wells.
3-3
-------
WDC 616?1 AO02
Y////////////////////,
..
xxxxxxxx/xxxxxxxxx/xxx
to
l\' Low Transmissivity-Xv
to
= X'x High Transmissivity Xv
Figure 3.1
COMPARISON OF THE CONE OF DEPRESSION IN LOW
CONDUCTIVITY AND HIGH CONDUCTIVITY AQUIFERS
-------
The Amphenol case study provides an example of a site where
aquifer heterogeneity led to differing patterns of flow over
the depth of a single aquifer. Vertical variations in
aquifer flow and contaminant distribution were studied by
monitoring multi-level wells. The resulting extraction
system included separate recovery wells for the shallow and
deep portions of the aquifer.
Problems associated with aquifer heterogeneity are
potentially most severe in fractured-rock or karst aquifers.
In these cases, the aquifer may not behave as a porous
medium, but may be more properly represented as a network of
more-or-less interconnected conduits. In such complex
systems, the design of aquifer remedial measures may be
reduced to trial and error. At the Utah Power & Light site,
for instance, some of the extraction wells that were
installed failed to produce water at high enough rates to be
useful. Other wells had to be installed to replace them,
and to a certain extent, the system operators have to live
with whatever production they can get from a well. It can
also be questioned at this site whether the potentiometric
head maps prepared on the basis of observed water-level
measurements provide an accurate description of the actual
ground-water flow patterns. In the absence of any other
technique, head maps are being relied on to verify that the
desired capture zones are being maintained.
Nonuniform distributions of aquifer properties make the
design of effective migration control systems much more
complicated. In these situations, the flow patterns that
will be generated by extraction wells are usually predicted
with the help of numerical models of ground-water flow.
Many computer programs are available for this purpose, and
some of them are quite elaborate. However, to get reliable
predictions using the models, a very thorough site
investigation may be required to get an adequate under-
standing of the spatial distribution of aquifer properties.
Even after intensive investigation there will always be a
certain degree of uncertainty about the correct numerical
representation of the aquifer. Calibration and verification
of the model against site data and testing of the model's
sensitivity to changes in parameters should be used to
determine appropriate factors of safety in the design of the
migration control system.
3.2 FACTORS AFFECTING AQUIFER RESTORATION
When the remedial goal is aquifer restoration, the time
required to clean up the contaminated aquifer becomes an
3-4
-------
important consideration. To achieve timely aquifer
restoration, the ground-water extraction system must not
only produce the desired flow directions but must also
generate the highest practical flow velocities throughout
the contaminated region. These hydrodynamic objectives are
generally necessary, but not sufficient, to assure the
success of the aquifer restoration.
In many cases, depending on the contaminant and aquifer
characteristics, the movement of large quantities of ground
water may not produce corresponding reductions of
contaminant concentrations. This can happen for several
reasons. The ground-water flow generated by the extraction
system may not penetrate into low-permeability zones in the
aquifer, the contaminants may be strongly sorbed to the
aquifer materials and therefore move at a lower velocity
than the ground water, or there may be residual sources that
replace the contaminants as they are removed from the
aquifer.
3.2.1 Well Placement and Pumping Rate
Because of the concern for timely performance in aquifer
restoration, the configuration of the extraction well
systems are likely to be different than they would be for
plume containment systems. For efficient contaminant
removal, it is necessary to minimize the time required for
contaminants to travel to the recovery wells.
One approach to this is to install a large number of
recovery wells distributed throughout the contaminant plume.
In low-transmissivity aquifers this may be necessary because
of the small radius of influence of the individual wells.
However, interference between wells in such a system causes
areas of low hydraulic gradient, known as stagnation
regions. In these stagnation regions there is practically
no ground-water flow, and contaminant removal may be very
slow. In some cases this problem can be overcome by
alternating the operation of adjacent wells. A more common
approach, however, is to progressively reduce the number of
operating wells by turning off peripheral wells as the size
of the contaminant plume is reduced. This method has been
used at the Fairchild Semiconductor and IBM-Dayton case
study sites.
In relatively transmissive aquifers, it may be more
efficient to concentrate the recovery effort near the center
of the contaminant plume. In this way the stagnation
regions generated can be reduced to small areas between
adjacent wells rather than broad areas between widely spaced
3-5
-------
wells. As much as possible, multiple recovery wells should
reinforce each other rather than draw ground water in
conflicting directions. The costs of piping the extracted
ground water to a central treatment plant can also be
reduced in this way. This was the approach used at the Des
Moines TCE site. Here, the aquifer was transmissive enough
that a cluster of 6 extraction wells located near the
contamination source could produce a capture zone
encompassing the entire contaminant plume.
The selection of well placement and pumping rates are
related. The zone of capture of a well network can be
increased both by increasing individual well capacity and by
increasing the number of recovery wells. The usefulness of
increasing pumping is limited in cases where drawdowns
become too great for efficient operation of the wells. In
these cases, greater numbers of recovery wells are needed.
Operational pumping rates can also be limited by the size of
the well casing and pump installed at each well location and
by interference between wells.
3.2.2 Contaminant Sorption and Retardation
Many contaminants tend to adsorb to and be retained by the
solid materials of the aquifer. The effect of sorption to
the stationary aquifer materials is to retard the rate of
contaminant migration and defeat the efforts of the aquifer
restoration system. The relative tendency for contaminants
to be retarded by adsorption depends on the chemical
properties of both the contaminants and the aquifer
materials. Contaminants that are highly soluble adsorb less
and are more mobile. Contaminants with low aqueous
solubility are more highly sorbed and less mobile. Soils
with high organic carbon content or high percentages of clay
mineral are more adsorbent and tend to retain contaminants
longer. The solubility, density and mobility of common
contaminants are shown in Appendix A.
The effect of contaminant sorption on aquifer restoration
systems can be considered in a couple of different ways. In
the analysis of solute transport in porous media, it is
customary to account for adsorption through a retardation
factor. (See Appendix B for the definition of the
retardation factor.) Using this concept, the migration
velocity of the adsorbing contaminant can be deduced by
dividing the interstitial ground-water flow velocity by the
retardation factor. Thus, if the time required for water
particles to travel from the hydraulically most remote edge
of the contaminant plume to the extraction wells can be
predicted, the aquifer cleanup time can also be estimated by
3-6
-------
multiplying by the retardation factor. Of course, this
estimate will be based only on advective contaminant
transport, neglecting hydrodynamic dispersion.
The advantage of this simplified estimation procedure is
that it requires only a hydrodynamic analysis of the flow
field generated by the extraction system. If the effects of
dispersion are to be included in the estimate, solute
transport modeling is required. This usually means that a
numerical model must be employed, because closed-form
solutions for solute transport are available for only the
simplest flow configurations. None of the detailed site
data collected in this investigation indicated that solute
transport modeling had been used to predict the remediation
time at any of the case study sites.
A second way of considering the effects of contaminant
adsorption is based on the concept of pore volume sweeping.
In this method, it is recognized that sorption results in
the partitioning of contaminants between the solid and
aqueous phases. This means that at any given time a certain
percentage of the contaminant mass is dissolved in the
flowing ground water, and the rest is adsorbed to the
stationary aquifer materials. If linear equilibrium
sorption is assumed, it can be shown (see Appendix B) that
the fraction of the total contaminant mass that is in
aqueous solution is the reciprocal of the retardation
factor. The restoration of the contaminated aquifer can
then be viewed as a process of sweeping fresh ground water
through the contaminant plume, allowing it to desorb
contaminants from the aquifer materials, and removing the
contaminated ground water through the extraction wells.
This concept was obviously adopted by several of the
designers of the aquifer remediation systems in the case
studies, because the expected progress of the remediation
was often expressed in terms of numbers of pore volumes
removed.
The use of the pore volume sweeping concept leads to the
expectation that the contaminant concentrations produced
from the recovery wells will decline exponentially with
time. Experience drawn from the restoration case studies
indicates that the concentrations typically do tend to
decline rapidly at first, and more slowly later. However,
the data are usually too irregular to conclude that an
exponential representation is accurate.
It has often been asserted that the linear equilibrium
sorption model is an over-simplification of the sorption-
desorption process and that kinetic constraints cause the
3-7
-------
desorption of contaminants to lag behind the rate at which
ground-water flow can remove contaminants from the
aquifer. The result of this phenomenon would be that
residual contamination remaining in the aquifer after the
cleanup was thought to be complete would cause ground-water
contaminant concentrations to rise again after the restora-
tion system was shut off. This has not been observed at any
of the case study sites in this investigation. System
operation has been terminated at only two of the case study
sites: Emerson Electric and IBM Dayton. At Emerson
Electric limited post-termination monitoring has given no
indication of a rise in concentrations. At IBM Dayton,
post-termination monitoring did indicate the re-emergence of
the contaminant plume, but this has been attributed to the
presence of contaminants in a residual nonaqueous phase
(NAPLs).
3.2.3 Isolation of Low Permeability Zones
Another factor that can impede the progress of aquifer
restoration systems is the presence of trapped contaminants
in low-permeability layers or lenses within the aquifer.
Ground-water extraction will cause preferential flow in the
high-permeability zones, bypassing the silt and clay lenses.
Contaminants in the low-permeability zones will then remain
as residual sources after the rest of the aquifer has been
cleaned up. This problem is thought to be impeding the
progress of the remediation at the Ponders Corner case study
site.
3.2.4 Contaminants in Nonaqueous Form (NAPLs)
Aquifer restoration at sites where residual contamination is
present in nonaqueous form is likely to be very slow. The
reason for this is that the mass of contamination contained
in the NAPLs is generally very large compared to the mass
that can be carried away in aqueous solution. Consequently,
a long time must pass before the flowing ground water can
exhaust the supply of contaminants contained in even a
relatively thin NAPL layer.
If the nonaqueous liquids are less dense than water and tend
to float on the water table, it may be possible to
accelerate the removal process by recovering some of the
contaminants as nonaqueous liquids. This is often done at
fuel spill sites. However, as the removal progresses, the
NAPL saturation of the soil will be reduced and the NAPLs
will cease to flow readily to recovery wells. This results
in an almost immobile NAPL-contaminated zone that still
contains large amounts of contamination. Auxiliary
3-8
-------
restoration techniques such as vapor extraction or bio-
remediation can sometimes be used to further accelerate the
removal of the floating NAPLs. At the Verona Well Field
case study site, vapor extraction is being used in
conjunction with ground-water extraction to deal with a
floating NAPL layer.
If the NAPLS are more dense than water, they can be expected
to sink through the aquifer, leaving a path of NAPLs at
near-residual saturation in their wake. If there is a
sufficient volume of the NAPLs, they may reach the bottom of
the aquifer and flow along its surface. In such cases, it
may be difficult to locate all of the NAPL pools, and the
technical means for removing them are limited. Several of
the case study sites are believed to have residual sources
of dense NAPLs. In each of these cases it has been
concluded that aquifer restoration is not feasible.
3.2.5 Leaching of Contaminants from the Vadose Zone
Contaminants adsorbed to the soils above the water table
represent a source of continuing aquifer contamination that
can significantly lengthen the time required for aquifer
restoration. Several approaches have been used in the case
studies to deal with vadose zone contamination.
At the Amphenol site, soil aeration was used as part of the
lagoon closure program before ground-water extraction was
started. The mass of volatile contaminants was thus reduced
to an allowable residual concentration that would not cause
leaching into the saturated part of the aquifer to interfere
with the aquifer restoration.
At the Ponders Corner site a vapor extraction system was
installed to reduce the contaminant content of the vadose
zone soils.
At the Utah Power & Light site the contaminant source area
was capped to reduce the rate of infiltration through the
contaminated vadose zone.
WDCR436/096.50
3-9
-------
Chapter 4
INFORMATION REQUIREMENTS FOR SYSTEM DESIGN AND OPERATION
The degree of site characterization necessary to design an
effective ground-water extraction system depends on the
objectives of the remediation. In cases where well-head
treatment is the objective, it may be sufficient simply to
characterize contamination to the extent necessary to design
the treatment system. The site characterization require-
ments for an aquifer restoration system are likely to be far
more extensive. Before designing an aquifer restoration
program, it is generally necessary to characterize: 1) the
hydrogeologic properties of the geologic layers involved,
2) the types and distributions of contaminants, 3) the
location of sources and their potential for continued
contamination of the saturated zone, and 4) the contaminant
migration properties of the contamination in the affected
aquifers. In addition, successful operation of the system
requires that the actual performance of the implemented
design be monitored so that adjustments can be made to
optimize performance.
4.1 HYDROGEOLOGIC INFORMATION
Whatever the remedial objective of the ground-water
extraction system, its functional purpose is to establish
some form of control over ground-water flow in the vicinity
of the wells. To design a system of wells that can
establish this control requires an adequate understanding of
the hydrogeologic characteristics of the site. For the
design of an extraction system, the necessary hydrogeologic
information includes the stratigraphy, the hydraulic
conductivity of aquifer layers, the leakance of semi-
confining layers, and the natural distribution of
potentiometric head.
4.1.1 Stratigraphy
It is necessary to identify the number and thickness of the
aquifers potentially affected by contamination at the site.
It is also important to know the lateral extent and
continuity of any confining or semi-confining layers. Any
local gaps in the confining layers that occur within the
area of concern can have an important effect on the success
of the aquifer remediation. This is illustrated at the IBM
Dayton site, where local holes in the clay layer between the
upper and lower aquifers are blamed for permitting DNAPLs to
enter the lower aquifer.
4-1
-------
It is also very important to establish the bottom of the
zone involved in the remediation. This is illustrated at
the Dupont Mobile site, where offsite flow of contaminants
in the lower portion of the aquifer is suspected as the
failure mechanism of the migration control system.
Adequate investigation of site stratigraphy requires
installation and appropriate logging of enough exploratory
borings to characterize the site. The number of borings
required depends on the complexity of the subsurface
configuration.
4.1.2 Aquifer Hydraulic Properties
In relatively homogeneous aquifers the main reason for
investigating the transmissivity is to predict the extent of
the capture zone that can be established by wells and the
rate of pumping that will be required. The most reliable
way of determining this is by conducting an aquifer test.
Aquifer test results were used in the design of the extrac-
tion systems at most of the sites covered in the case
studies. The exceptions to this are Site A, and the Emerson
Electric site. At Site A, the design was apparently based
on regional hydrogeologic information. At Emerson Electric,
the design was based on slug testing.
When the stratigraphy is well defined and good trans-
missivity estimates can be made, the depth-averaged
hydraulic conductivity can easily be determined. Good
hydraulic conductivity estimates are necessary if the
velocity of the contaminant migration induced by the
extraction wells is to be estimated. However, the depth-
averaged hydraulic conductivity estimates arising from
aquifer test results may be deceptive if the aquifer is
vertically heterogeneous. Determination of vertical
hydraulic conductivity variations within an aquifer unit
requires special testing methods. In open hole wells this
can be done by packer testing or spinner logging. In
unconsolidated materials, individual slug tests on wells
screened at different depths or comparative of grain size
analyses can be used to estimate vertical permeability
variations.
At the Nichols Engineering site multiple aquifer tests were
run at different depths in the fractured bedrock to develop
a conceptual understanding of vertical variations in
permeability.
In aquifers with complex heterogeneity, conventional aquifer
tests can be difficult to interpret. A more useful approach
4-2
-------
Chen is to run a long-term aquifer test in which a large
number of observation wells is monitored. An example of
this is the aquifer test done at the Amphenol site, where
all of the site monitoring wells were used to observe the
effects of a 3-day pumping test. The resulting flow
patterns were then used to guide the design of the ground-
water extraction system.
Another useful way of dealing with complex hydrogeologic
systems is pilot testing of extraction wells. This usually
involves incremental design of the extraction system as
successive components are installed and tested. This
approach was used at most of the larger case study sites.
The determination of aquifer storage coefficients is of
secondary importance at most sites because the operation of
the extraction system is usually analyzed as a steady-state
phenomenon. Normally, the time required for significant
contaminant migration to occur is much longer than the
transient response period of the aquifer.
In multi-layered hydrogeologic systems, the thickness and
hydraulic conductivity of the semi-confining layers
separating individual aquifers can have an important
influence on the effectiveness of ground-water extraction
systems. Inter-aquifer leakage in response to the pumping
of extraction wells can drastically reduce the radius of
influence of the wells. This may be what is happening at
Site A, where pumpage of the extraction well has not
produced a noticeable drawdown in any of the monitoring
wells.
Leakage through semi-confining layers also permits
contaminants to move between aquifers. This was a major
consideration at the General Mills and Utah Power & Light
sites, where the extraction systems in each aquifer were
designed to reverse natural downward flow in the areas of
high contamination.
4.1.3 Potentiometric Gradients
It must be recognized that ground-water extraction systems
achieve their remedial goals through the manipulation of
ground-water flow patterns in the contaminated aquifers. At
most sites, the potentiometric gradients produced by the
extraction wells must compete with a larger scale regional
gradient that is attempting to move the contaminants in an
undesired direction. These regional gradients may arise
naturally, or they may be the result of nearby production
wells that are to be protected from contamination. In any
4-3
-------
case, the design of the extraction system must take them
into account. This requires the installation and monitoring
of a sufficient number of piezometers or water level
monitoring wells. Since natural gradients can change
seasonally, the water levels in the wells must be measured
enough times to determine the range of gradients that may
occur. The Amphenol and Utah Power & Light case studies
illustrate the importance of seasonal variations in
potentiometric head.
At most of the case study sites the pre-existing gradients
were taken into account in the design of the extraction
systems. At Site A, the horizontal gradients were said to
be small and were effectively neglected in the design of the
extraction system. Vertical gradients at Site A were
apparently neglected also.
4.2 CONTAMINANT DISTRIBUTION AND CHARACTERISTICS
Before an effective remedial response can be designed, it is
necessary to evaluate the nature of the ground-water
contamination problem. This involves determination of the
number and identity of the contaminant compounds present,
their concentration and spatial distribution in the aquifer,
and their mobility characteristics.
4.2.1 Identification of Contaminants
This is probably the aspect of the site investigation that
is routinely most thoroughly covered. It is important in
the design of an extraction system because of the different
mobility and toxicity of the various contaminants and
because of the special handling and treatment problems
associated with some compounds.
At most of the sites presented in the case studies, the
contaminants of concern were volatile organics. These
compounds have relatively high mobility as dissolved species
in ground water. They are also relatively easy to treat
using air stripping or carbon adsorption.
When a variety of compounds with different characteristics
is present, the design of the recovery system may become
more complicated due to the different mobilities involved.
At the Ville Mercier site the ground water was contaminated
with a large number of different compounds. Since the
extraction system objective was migration control, the wide
range of mobilities had no effect on the placement of the
wells. However, if the intent were to remediate the aquifer
4-4
-------
expeditiously, multiple extraction systems might be required
to address individual groups of compounds that had travelled
different distances from the source.
In addition to the chemical identity of the compounds, it is
also important to determine whether they are present in a
nonaqueous phase. If NAPL contamination is involved, it is
unlikely that aquifer restoration by an extraction system
will be successful. At several of the case study sites,
NAPL contamination was known to be present. At the IBM
Dayton, Utah Power & Light, and Ville Mercier sites, the
remedial objective was limited to migration control because
of them. At the IBM San Jose site slurry wall containment
was used to isolate the NAPL-contaminated region, and at
Verona Well Field a vapor extraction system was installed to
deal with it.
At several of the other sites, no special site investigation
was conducted to determine the presence of NAPL contamina-
tion, even though the mode of waste disposal or the high
concentrations in the ground water might have suggested it.
Because of interfacial tension effects, NAPLs that are
present in an aquifer at relatively low saturation may not
flow into a well. Therefore, the presence of NAPEs cannot
be ruled out just because they have not been seen as a
separate phase in any of the ground-water samples. A more
reliable way to detect nonaqueous phase contamination is to
take soil samples during installation of the monitoring
wells (e.g. split-spoon or Shelby Tube samples) or as part
of a separate soil boring program. In fractured rock
aquifers, rock coring can be used to collect aquifer samples
but this may not conclusively rule out NAPL presence.
4.2.2 Contaminant Distribution and Concentration
Knowledge of the spatial distribution of contaminants in the
aquifer is obviously very important in the design of an
extraction system because it determines the area in which
control of ground-water flow must be established. The
problem is, essentially, to determine the boundaries of the
contaminant plume. This generally requires monitoring wells
to-be installed both inside and outside the contaminated
area. If all of the wells are inside the contaminated area,
its edge cannot be located.
In the design of plume containment and well-head treatment
systems it may not be essential to establish the upgradient
extent of the contaminant plume. But, the downgradient and
lateral extent of the area in which health-based standards
are exceeded must be known.
4-5
-------
Accurate measurement of contaminant concentrations is
important for several reasons. First, the aquifer
restoration goals are expressed in terms of individual
contaminant concentrations, so the boundaries of the
contaminated region are determined by the concentrations of
the contaminants. Second, predictions of the contaminant
concentrations in the extracted ground-water are required in
the design of the treatment system. Third, the best way to
measure the progress of aquifer remediation is by periodic
comparison of the contaminant concentration distribution
with the initial distribution that existed before the
remediation began.
Concentration distributions are usually described by drawing
contour maps of the contaminant plume. In most of the case
studies, contaminant contour maps are presented for the
initial site conditions and for subsequent times during the
remediation period. This greatly facilitates the evaluation
of remedial effectiveness. However, no concentration
contour maps were available for the Emerson Electric,
Site A, and Utah Power & Light studies.
When interpreting the meaning of concentration contour maps
it is important to understand how vertical variations in
contaminant concentration have been accounted for. In many
cases, the possibility of vertical concentration gradients
is neglected. The contour maps may be developed directly
from the analytical results without regard to the depth from
which the samples were taken or the length of the screened
intervals in the monitoring wells. This can give a false
impression of the distribution of contaminants in the
aquifer. If the samples were taken from fully penetrating
monitoring wells, the results can be interpreted as a
permeability-weighted depth-averaged concentration. Such a
concentration measurement is useful in predicting the
contaminant concentrations that will be produced by fully
penetrating extraction wells. However, concentrations
measured in this way will probably be lower than the maximum
concentrations in the aquifer and may produce incorrect
estimates of total contaminant mass.
Estimates of the total contaminant mass in the aquifer are
sometimes used as an expression of the magnitude of the
problem to be remediated. At the Savannah River Plant, the
goal of the aquifer restoration system was to remove 99% of
the contaminant mass in the aquifer in 30 years. Evaluation
of the performance of this system therefore depends heavily
on the accurate estimation of this contaminant mass. This
requires an understanding of the three-dimensional
distribution of contaminant concentration in the aquifer,
4-6
-------
which can only be obtained by multi-level sampling. Both
dissolved and sorbed contaminants must be accounted for in
estimates of the total contaminant mass (see Appendix B).
4.2.3 Contaminant Mobility Characteristics
One of the most common explanations for the underestimation
of the time required for aquifer remediation is that the
effects of adsorptive retardation were not accounted for.
This may have been the case at Site A and at the General
Mills site. Adsorptive effects may also have been under-
estimated at the Savannah River Plant. At most of the other
sites studied, the retarding effects of contaminant sorption
were acknowledged by estimating that multiple pore volumes
would have to be extracted before the concentrations could
be reduced to the regulatory standards.
Quantitative estimates of contaminant retardation are
usually based on the total organic carbon content of the
aquifer materials and tabulated values of partition
coefficients for organic compounds. At most of the case
study sites, however, no measurements of total organic
carbon in the soils were made.
At the Amphenol site, soil samples were taken from beneath
the contaminant source area and subjected to laboratory
testing to evaluate the partitioning of contaminants between
soil and water. This was done to support analysis of the
continued leaching of contaminants from the contaminated
vadose zone into the underlying aquifer. Laboratory testing
of site materials to determine partition coefficients is
fairly rare at ground-water contamination sites, but the
information obtained from such tests is quite useful.
4.2.4 Identification of Contaminant Sources
Adequate characterization of the sources of contamination is
important in the design of most aquifer remediation systems.
This is particularly true of systems designed for aquifer
restoration, where the existence of continued contaminant
input from residual sources can defeat the ground-water
cleanup efforts.
Usually, the original source of the contamination is known,
although this was not the case at the Site A and Emerson
Electric sites. But, when the original source has been
removed, it is important to evaluate the potential for
continued leaching of contaminants from the remaining
contaminated soils. It may be advisable to consider
4-7
-------
additional soil removal or surface capping to reduce
infiltration of contaminants to the ground water.
4.3 PERFORMANCE MONITORING REQUIREMENTS
The performance of the ground-water extraction system must
be monitored regularly to ensure that the desired control is
being maintained over the ground-water flow patterns and the
movement of contaminants.
Monitoring of the hydraulic performance of the system is
done by regular measurement of water levels in piezometers
and monitoring wells throughout the area of remediation.
Potentiometric surface maps are then drawn for each of the
aquifers of concern to show that the desired capture zones
are being maintained. In multi-aquifer situations the
potentiometric gradients between aquifers are also checked
using these maps.
Monitoring of the aquifer cleanup effectiveness of the
extraction system can be done in several ways. One way that
is rather common, but not very reliable, is to monitor the
flow rate and contaminant concentrations produced by the
extraction wells. Integration of the product of flow and
concentration over time gives an estimate of the mass of
contaminants removed. This is an interesting statistic, but
it does not provide a direct measure of the reduction in
contaminant concentrations in the aquifer.
The measurement of contaminant concentrations should not be
limited to samples taken from the production wells. Samples
should also be taken simultaneously from enough monitoring
wells in the remediation area to permit plume maps to be
drawn. It is such plume maps that should ultimately be used
to determine when the aquifer remediation is complete.
Samples taken from the extraction wells are not reliable for
this because of the considerable dilution that these wells
generally produce through their hydraulic effects on the
aquifer.
In monitoring plume containment systems, it is important
that some monitoring wells be located downgradient of the
extraction wells. This is necessary so that flow reversal
downgradient of the well can be demonstrated. Hydraulic
measurements are usually relied on as the primary indicator
of success in plume containment systems. However, periodic
water quality sampling should be performed also to ensure
that contaminants are not escaping by passing under or
between the wells. As illustrated by the Dupont Mobile
4-8
-------
site, the appearance of complete hydraulic capture on the
basis of potentiometric heads can be deceptive if the
hydrogeologic nature of the aquifer is not well understood.
4.4 POST-TERMINATION MONITORING
Performance monitoring at the aquifer restoration sites
should continue even after the extraction system ceases to
operate. It is to be expected that the low contaminant
concentrations measured toward the end of the remedial
action may rebound after the extraction system has been
turned off. This can happen for a variety of reasons. One
possibility is that the ground-water flow patterns generated
by the extraction system can cause dilution of the concen-
trations sampled at monitoring wells. Another possibility
is that residual contaminants stored in low permeability
zones of the aquifer sorbed to the aquifer materials, or
retained as NAPLs, may cause the concentrations to rise when
the extraction is terminated. Also, the recovery of water
levels after system shutdown may resaturate contaminated
soils.
This phenomenon has been observed in only one of the case
studies, at the IBM Dayton site. At Emerson Electric, two
rounds of post-termination sampling from the extraction
wells have not indicated any resurgence of contamination.
None of the other sites have progressed far enough for the
extraction systems to be shut down.
WDCR436/098.50
4-9
-------
Chapter 5
METHODS OF INCREASING SYSTEM EFFECTIVENESS
Several enhancement techniques have been proposed as ways to
improve the performance of ground-water extraction systems,
either by accelerating the aquifer remediation or making it
more efficient. At most of the case study sites in this
investigation, conventional ground-water extraction,
consisting of continuous pumping from a fixed number of
recovery wells was the only technology employed. However,
at a few of the sites attempts were made to increase the
effectiveness of the basic system. The enhancement
techniques used included: progressive modification of the
extraction well configuration, pulsed pumping, auxiliary
physical containment, reinjection, vapor extraction, and
fracture enhancement.
5.1 PROGRESSIVE SYSTEM MODIFICATION
At sites where the ground-water contamination covers a broad
area it is often necessary to initiate ground-water
extraction using numerous widely spaced wells. As the
remediation progresses, the outer edges of the contaminant
plume will probably be cleaned up first, and the wells that
have been located there will become useless. Furthermore,
continued operation of these wells will interfere with the
progress of the cleanup in the areas that still have
contamination. It is, therefore, necessary to turn off
these outer wells, and perhaps to increase the rate of
withdrawal in the interior of the remaining plume.
An example of this is the Fairchild Semiconductor site,
where as many as 14 offsite wells were pumped initially. As
the remediation has progressed, the number of offsite
extraction wells pumping has been reduced to 5.
Another example is the IBM Dayton site, where up to 17
extraction wells have been used simultaneously to reduce the
extent of the contaminant plume leaving the site. Past
experience has shown that these wells can probably reduce
the size of the plume so that it can be confined within the
bounds of the IBM property. Accordingly, a phased reduction
in the number of extraction wells is planned so that
eventually only one or two onsite wells will be sufficient
to control contaminant migration.
5-1
-------
5.2 PULSED PUMPING
Pulsed pumping is a technique that has been proposed as a
way to increase the efficiency of ground-water extraction
systems in situations where the slow release of contaminants
from residual sources within the aquifer is controlling the
rate of cleanup. These residual sources may be contaminants
adsorbed to the aquifer materials or contained in low
permeability lenses, or they may arise from the slow
dissolution of contaminants from residual NAPLs.
The idea behind the pulsed pumping concept is that, since
the rate of mass removal is primarily controlled by the
release of residual contaminants rather than by the velocity
of ground- water flow, it does no good simply to pump more
ground water. By pumping intermittently the ground water
passing through the residual source region can be allowed to
dissolve or desorb contaminants from it almost to equili-
brium while the recovery wells are off. Then, when
extraction resumes, the ground-water produced will carry out
a higher load of contamination.
None of the case study sites in this investigation have
implemented a formal plan for pulsed pumping. At the GenRad
site, the extraction wells were turned off in the winter to
prevent freezing of the above-ground piping. This might
have been expected to result in higher contaminant
concentrations in the extracted water when the wells resumed
pumping in the spring. Unfortunately, the quarterly
sampling frequency used at this site is too low for such a
phenomenon to be observed.
At the Utah Power & Light site, the individual wells have
been operated intermittently, with production shifting from
one well to another. This has been done partly because the
productivity of the shallow aquifer is reduced during the
winter and spring months by low water levels. In this case,
also, the data collection frequency has been too low to
determine whether there is any correlation between
interrupted pumping and the concentrations produced.
5.3 PHYSICAL CONTAINMENT SYSTEMS
At the Fairchild Semiconductor site a slurry wall was
constructed around the Fairchild property in the uppermost
contaminated aquifer. The purpose of the wall was to
prevent offsite movement of contaminants both by physical
intervention in the flow system and by facilitating the
reversal of gradients across the site boundaries. As a
5-2
-------
result of ground-water extraction inside the slurry wall,
the onsite portion of the uppermost aquifer has been
dewatered. This has helped to reverse the vertical
gradients and prevent the flow of contaminated ground-water
to underlying aquifers. It has also prevented flushing of
the aquifer materials inside the slurry wall and the
collection of ground-water samples there.
5.4 REINJECTION
Reinjection of treated ground water to the aquifer is
frequently considered as a means of increasing the velocity
of ground-water flow toward the recovery wells. It would
also serve as a disposal method for the treated ground
water.
Reinjection was used briefly at the IBM Dayton site, where a
line of offsite injection wells was intended to increase the
rate at which the contaminant plume was drawn back to the
site. However, the injection wells rapidly became clogged
and had to be dropped from the system.
Reinjection was also considered at the Utah Power & Light
site. It was never implemented, however, because the high
potentiometric head that would be produced around the
injection wells might induce unwanted vertical flow into the
lower aquifers.
5.5 VAPOR EXTRACTION
Vapor extraction was used as an enhancement technology at
the Ponders Corner and Verona Well Field sites. At Ponders
Corner the vapor extraction system was installed at the
original contaminant source to remove adsorbed contamination
in the vadose zone. No information has been provided on the
performance of this vapor extraction system. However, if
the main impediment to aquifer cleanup at the Ponders Corner
site is the slow release of contaminants from low
permeability zones in the saturated part of the aquifer, the
effectiveness of the vapor extraction system in speeding up
aquifer remediation may be limited.
At the Verona Well Field site, a vapor extraction system has
been installed to help in the removal of contaminants from a
floating NAPL layer. The vapor extraction system is
positioned directly above the ground-water extraction system
so that they can both work in concert. In the 22 months
between March 1987 and January 1989, the ground-water
5-3
-------
extraction system removed an estimated 11,000 pounds of
volatile organic compounds (VOCs). During the 10 months
between March 1988 and January 1989, the vapor extraction
system removed approximately 27,000 pounds of VOCs.
Obviously, the vapor extraction system is helping
significantly in the removal of volatile contaminants.
At sites with floating layers of volatile contaminants,
ground-water and vapor extraction systems can potentially be
used together to great advantage. The drawdown produced by
the ground-water extraction system thickens the vadose zone
so that the vapor extraction system can reach deeper into
the aquifer. The vapor extraction system, in turn, .reduces
the atmospheric pressure above the water table and permits
higher ground-water withdrawal rates with the same drawdown
because of increased pressure gradient. In addition, if the
systems are operated to produce a fluctuating water table,
the NAPL saturation of the soil can be kept low. This
increases the interfacial area of contact between the NAPLs
and the moving water and vapor phases so that the rates of
volatilization and dissolution will be increased. It also
increases the air and water permeability of the soil in the
NAPL-contaminated zone so that contaminant removal can be
accelerated. At the Verona Well Field site the tandem
operation of the ground-water and vapor extraction systems
has not as yet been manipulated to maximize these
possibilities.
5.6 FRACTURE ENHANCEMENT
The Black & Decker case study site involved contamination of
a fractured rock aquifer in which the fracture density was
relatively low. Because of the low fracture density, efforts
to establish a hydraulic capture zone in the aquifer by
pumping wells were unsuccessful. In order to increase the
permeability and interconnectedness of the aquifer,
explosives were used to create more thorough fracturing in
the vicinity of the ground- water extraction wells. As
described in the Black & Decker case study, this technique
appears to have made possible the establishment of a capture
zone to control contaminant migration in the aquifer.
WDCR437/046.50
5-4
-------
WDCR437/050.50
APPENDIX A
DENSITIES, SOLUBILITIES, AND MOBILITY CHARACTERISTICS OF ORGANIC HAZARDOUS
SUBSTANCES FREQUENTLY FOUND AT PROPOSED AND FINAL NPL SITES*
Chen leal Name
1,1,2-Trlchloropthylene (TCE)
Toluene
Benzene
Chloroform .
rrB-i?6n/rcB-i254
1,1, 1-Trltrhloroethane
Tpt rach 1 oron thenp
Phenol
Ethylbpnzene
Xylpnp (meta-, ortho-, para-)
1,2-Trans-DlchloroPthylpne
flpthylpne Chloride
1,1-Dlchloroethane
1 , 1-rtchloroethene
Vinyl Chlorlilp
Chlorohenzene
Carbon Tetrachloride
l,2-Dtchloroeth«ne
Ppntachlorophenol (PCP)
Naphthalene
Methyl Ethyl Kptone
Acetone
Phenanthrene
Bpnzo (a) Pyrene
1,1, 2-Trlchloroethane
DDT
Anthracenp
BHC Gamma
B!s(2-ethylhexyl)Phthalate
1,1,2, 2-Tet rach 1 oroethane
Styrene
Benzol J ,k) Fluorene
Pyrene
1 ,2-Cls-Dlchloropthylene
Fluorpnp, NOS
Tr 1 rhlorof luorone thane
t'1-N-Butyl Phthlntn
Oilorilnne
A< pnnpthene
Ithyl Chlnrldp
Frequency0
331
259
218
182
170
163
155
124
120
119
105
95
89
88
79
69
65
65
59
52
43
33
29
27
25
24
24
23
23
22
22
20
19
19
17
15
15
15
15
15
Density0
1.46 - 1.49
0.866
0.1787
1.474 - 1.478
1.4 - 1.5
1.3376
1.58658
1.071
0.866
0.86
1.28
1.307 - 1.361
1.1P80 - 1.1757
1.2129
1.33 - 1.39
1.107
1.589
1.2569
1.978
0.9628 - 1.162
0.805
0.788
1.179
1.4416
1.241 - 1.306
1.25
0
0.981
1.58658
0.0059
.271
.28
.202
.494
.048
l.S<) - 1.61
1.189
O."214
Reference
1
1
1
1
3
1
1
1
1
3
1
1
1
2
I
1
1
1
1
1
1
1
1
2
1
—
2
1
1
1
1
1
3
1
1
1
Solubility
0 tag/I)
l.lxlO3
5.067x10,
1.787x10,
8.216x10
7x10
1.554xl03
1.49x10*
9.3x10,
, 1.53x10*. ,
1.46 x 10*. 2.13 x 10 , 1.85 x 10
3.9x10?
1.37x10,
5.5x10*
3.2x10
1.1
4.72x10?
8.00x10,
8.69x10,
1.4x10
3.1x10
2.68x10;!
2.3xl06
1.29.J
4.9x10 *
4.5x10
S.OxlO"3
l!sxlO~*
3.14x10:
4.0x10 ,
2.9x10
3.2xl02
3X10"5 .
1.35x10,
8.0x10
1.98
1.10x10?
1 . 1x10
1.85
3.7
5.74x10
Reference0
4
5
6
7
8
7
7
9
6
10, 11
12
7
13
14
• 15
16
17
18
19
20
21
12
22
12
18
13
13
12
13
13
10
12
23
15
24
25
26
27
12
13
Q*
152
242
97
134
349,462/63,914
155
318
27
622
588, 363, 552
39
25
45
217
22 - 704
318
232
36
900
1,300
235
1
23,000
282,185
49
238,000
26,000
28,900
12,200
88
380
19,800
63,400
124
5,835
159
217
53,200
2,580
42
A
Mobility Class
Moderate
Moderate
High
Very High
luoblle
Moderate
Moderate
Very High
Low
Lov, Moderate, Low
Very High
Very High
Very High
Moderate
Very High-Low
Moderate
Moderate
Very High
Low
Low
Moderate
Very High
luoblle
Inoblle
Very High
Inoblle
Inoblle
Inoblle
Slight
High
Moderate
Slight
Immobile
High
Slight
Moderate
Moderate
Imnohlle
Slight
Very High
TliI'; tnh 1> Is for illustrative purposes only.
P.il ,t tin A vat InMe In selected ref rrenrps .
Sour'-*»: HUM HILL. Technical Support Document: Revised Hazanl Ranking System, FPA Contract 68-O7-7090.
^otirro; CH2M HILL. May IQS6. Ppwe<1l*l Investlqatlon, ?py;ponr Recycling Corporation, Volume 2. EPA Contract 68-Dl-€6Q2
-------
Appendix A
(Continued)
references: (complete reference citations are listed In Section 3.6)
1 - The Merck Index (Merck, 1976)
2 - Aldrlch Catalog/Handbook of Fine Chemicals (Aldrlch, 1983)
3 - U.S. Environmental Protection Agency Trratablllty Manual (U.S. EPA, 1981)
4 - Pogers R.D. and J.C. McFarlane. 1981. Sorptlon of Carbon Tetrachlorlde, Ethylene Dlbromlde and Trichloropthylene In Soil and Clay. Environ. Honlt. Assess. 1:155-8.
5 - Ross S.S. and W.H. Thomas. 1981. Solubility Behavior of Three Aromatic Hydrocarbons In Distilled Hater. Environ. Scl. Technol. 15:715-6.
r, - Chlou C.T., P.E. Porter, and D.H. Schmf-rtrtlna. ]O83. Partition Efrulllbrln of Nonlonlc Organic Compounds Between Soil Organic Matter and Hater. Environ. Scl. Technol.
17:227-31.
7 - llorvath A.L. 19B2. Halogenated Hydrocarbons: Solublllty-Mlsclblllty with Hater. Marcel Dekker, Inc., New York.
8 - Griffin R.A. and S.F.J. Chou. 1981. Movement of PCRs and Other Persistent Compounds Through Soil. Hater Scl. Technol. 13:1153-63.
1 - Callahan M.A., M.H. Sllmak, N.H. Gabel, I.P. May, C.F. Fowler, J.R. Freed, P. Jennings, P.L. Durfee, F.C. Hhltinore, and B. Maestri, et al. 1979. Hater-related
Environmental Fate of 129 Priority Pollutants. Volume II. EPA-440/4-79-029B. U.S. EPA, Hashlngton, D.C.
10 - National Academy of Sciences. 1980. The Alkyl Bentenes. Hashlngton, D.C.: National Academy Press. U.S. EPA Contract 6B-01-4655.
11 - Polak, J. and B.C.Y. Lu. 1973. Mutual Solubilities of Hydrocarbons and Hater at 0 and 25C. Can. J. Chen. 51:4018-33.
1? - General Sciences Corporation. 1986. CHEMEST and FAP Estimation Programs. Graphical Exposure Modeling System. Landover, MD: General Sciences Corporation. Prepared
for U.S. Environmental Protection Agency, Office of Toxic Substances. Contract No. 68-02-3970.
13 - U.S. EPA. 1982. Aquatic Fate Process Data for Organic Priority Pollutants
14 - Versar, Inc. 1983. Chemicals Selected for Use In a Detailed Statistical Analysis of CHD1EST. Prepared for Office of Toxic Substances, U.S. .Environmental Protection
Agency.
15 - Verchueren K. 1983. Handbook of Environmental Data on Organic Chemicals. 2nd ed. Van Hostrand Relnhold Co., Inc., New York.
16 - MacKay D., A. Bobra, H.Y. Shlu, and S.H. Yalkowsky. 1980. Relationships between Aqueous Solubility and Octanol-water Partition Coefficients. Chemosphere. 9:701-11.
17 - Deshon H.D. 1979. .Cartoon Tetrachlorlde. Klrk-Othner Encycl. Che». Tech. 3rd ed. 5:704-14.
in - Hllson J.T., C.G. Enfleld, H.,1. Dun lap, R.I.. Cosby, D.A. Foster, and L.R. Baskln. 1981. Transport and Fate of Selected Organic Pollutants In a Sandy Soil. J^
Environ. Oual. 10:501-6.
11 - KlUrr F.., I. Scheunert, II. Oeyrr, H. Klein, and F. Korto. 1979. Laboratory Screening of the VolatlllrMlon Rates of Organic Chemicals from Hater and Soil.
Chemosjphere. 8:751-61.
70 - Poarlman R.S., S.H. Yalkowskl, and S. Banerjee. 1984. Hater Solubilities of Polynuclear Aromatic and Heteroaromat Ic Compounds. J. Chem. Ref. Data. 13:*55-61.
21 - Papa A.J. and P.O. Sherman Jr. 1981. Ketones. Klrk-Othner Encycl. Chen. Tech. 3rd ed. 13:894-941.
22 - Karlctikhoff S.H., D.S. Brown, and T.A. Scott. 1979. Sorptlon of Hydrophoblc Pollutants on Natural Sediments. Hater Res. 13:241-8.
?1 - Means J.C., S.G. Hood, ,T.J. Hassett, and H.I,. Banwart. 1980. Sorptlon of Polynuclear Aromatic Hydrocarbons by Sediment and Soils. Environmental Scl. Technol.
14:1524-8.
24 - U.S. FPA. 1981. Treatablllty Manual I. Treatablllty Data. EPA-600/2-82-001A. U.S. EPA, Hashlngton, D.C.
25 - Smart B.E. 1980. Klrk-Othmer, 3rd ed., Vol. 10, pp. 856-970.
26 - Holfe N.L., H.C. Steen, and L.A. Burns. 1980. Phthalate Ester Hydrolysis: Linear Free Energy Relationships. Chemosphere. 9:403-8.
?7 - Hell L., G. Dure, and K.E. Ouentln. 1974. Solubility In Hater of Insecticide Chlorinated Hydrocarbons and Poljchlorlnated Blphenyls Jn View of Hater Pollution. Z^
Hasser Abwasser Forsch. 7:169-75.
-------
APPENDIX B
CONTAMINANT PARTITIONING IN SORBING MEDIA
One error commonly made in estimating the initial mass of
contaminants present in the saturated zone is to neglect the
contaminants sorbed to the solid phase. The contaminants sorbed
to the solid phase can recontaminate the liquid phase once
remediation has stopped, or they can slow the progress of
remediation. The total mass of contaminants present in the
saturated zone is given by:
MT = ML + Ms = RCnV (1)
where Mj = the total mass of contaminants in the saturated zone,
ML = the mass of contaminants in the liquid phase, MS = the mass
of contaminants sorbed to the solid phase, n = the porosity, C =
the concentration in the liquid phase in mass of contaminants per
volume of water, V = the bulk aquifer volume, and R = the
retardation factor. Equation (1) states that the mass of
contaminants in the saturated zone is equal to the retardation
factor times the mass contained in the liquid phase.
Derivation
The retardation factor, R, is defined as:
R = V* = 1 + Kd Db (2)
Vc n
where Vs = the average linear velocity of the ground water Vc =
the average linear velocity of the contaminants, Db= bulk soil
density, Kd= distribution coefficient.
The distribution coefficient, Kd, is defined as:
Kd= S/C, (3)
where S = mass of sorbed contaminants per mass of soil.
The mass of contaminants in the soil, M^ is given by:
Ms= S Db V, (4)
and the mass of contaminant in the liquid phase, ML, is:
ML= C n V. (5)
WDCR437/050.50/2
-------
Equations (4) and (5) can be combined to give the total mass of
contaminants in the two phases:
MT = ML + M, = S Db V + C n V (6)
If equation (3) is solved for S, then equation (6) becomes:
Mr = < Kd Db + n) C V, (7)
which, after rearrangement, becomes:
MT = (1 + K,, Db/n) C n V (8)
Substitution of equation (2) in equation (8) yields equation (1).
As an example, if the retardation factor in a given soil for a
given contaminant is 4, then the mass of contaminants present in
the solid phase is 3 times the mass of contaminants present in
the liquid phase. That is, the mass of contaminants present in
the liquid phase is given by equation (5), and the mass of
contaminants sorbed to the solid phase is given by:
MS = (R-l) C n V
WDCR436/097.50
&U.S. GOVERNMENT PRINTING OFFICE: MM - 74S-B9/M9M
------- |